Coined nearly 10 years ago by Patrick Debois, the term “DevOps” is now accepted to describe a methodology by which software development teams and IT operations teams work collaboratively for a continuous delivery approach. Here, leading IT executives offer their DevOps predictions for 2019.
As AI use increases, data science teams will adopt DevOps best practices: Companies that have implemented DevOps have seen increased business efficiency and faster deployments. Consequently, we foresee that, in 2019, as the demand for AI-driven applications continues to rise, that data science teams will adopt DevOps best practices in their model management workflows. A DevOps approach helps data scientists build automated pipelines to re-train, re-select, and re-deploy production models in a more stable way, while providing the ability to test multiple models deployed into production. This trend will accelerate as data science and application development teams begin working closer together to improve the efficiency of developing, deploying, and maintaining AI & ML-driven applications to meet demand across the enterprise. — Justin Charness, principal product manager, machine learning, Oracle
DevOps will focus more on integration at the edge between services: The traditional, on-premises database infrastructure model is clearly changing. Over the last few years, we have seen accelerating migration toward infrastructure as a service (IaaS), which includes the database environment in database as a service (DBaaS). The increased dominance of cloud services from AWS, Microsoft Azure, and Google Cloud—along with developing technologies, such as Kubernetes and the serverless model—are evidence that the trend will continue. With this trend, DevOps will focus on the integration at the edge between services. Much of the typical infrastructure operations work will be handled in software, with systems such as Kubernetes operators. With the increased acceptance that public clouds can meet regulatory requirements, including HIPAA and HITRUST, even traditionally premise-bound large enterprises are going to start shifting to cloud-based operations and infrastructure in order to realize the flexibility, agility, and cost-savings that IaaS offers. Enterprises will continue to utilize a combination of configuration management and container technologies to fully automate infrastructure management—which means the role of DBA will change from infrastructure management to strategic application deployment. — Peter Zaitsev, co-founder and CEO, Percona
There will be a deepened focus on functions: In today’s DevOps environments, technology professionals who have mastered operating containerized workloads in complex ways are working to streamline and optimize delivering these capabilities by leveraging functions as a service. The breadth and depth of this focus on functions will likely deepen throughout the next year, as more technology professionals become comfortable leveraging containers in production, and recognize benefits achieved through serverless computing—such as faster start-up times, better resource utilization, and finer-grained management. However, even with these recognizable benefits, future DevOps pros will become adept at determining the use cases where functions as a service and serverless computing are appropriate for their environments and resources. Without this acquired skill, companies that dive right into functions-as-a-service without understanding the benefits and pitfalls of running numerous individual functions for different tasks may see bigger bills as the result, and the tech pro trying to explain these bills to management and business leaders may see bigger problems. — Keith Kuchler, VP of engineering for SolarWinds Cloud
Automation remains key: Automated provisioning and lifecycle operations for bare metal infrastructure can help enterprises deliver on-premise services faster and reduce operational costs than public cloud. This is especially true for managing hybrid cloud environments. The hybrid cloud vision is resonating, but many customers are still too early in their journey to have achieved the results they have planned for. That is why we created the composable infrastructure category with the launch of HPE Synergy 3 years ago. With a single line of code, the solution's composable API can fully describe and provision the infrastructure that is required for applications, transforming IT operations while simplifying the entire DevOps process. Lastly, Kubernetes won the container orchestration war. In 2019, watch how everyone will be fleshing out their offerings. It will be interesting to see what the mix is for vendor supported versus 'roll your own open source,' and who grabs market share among the vendor offerings. — Said Syed, director of HPE UX Design and Developer Productivity, HPE Chief Design Office
Experimentation with continuous delivery approaches will expand: Adoption of the continuous delivery engineering approach and use of container related-technologies (such as Docker and Kubernetes) in large organizations will dramatically increase with the adoption of microservices and multi-cloud architectures. In addition, fully automated continuous deployment (leaving the entire chain of continuous integration (CI), continuous delivery (CD), and continuous deployment on autopilot) will get traction by companies embracing immutable infrastructure approaches and technologies (such as Spinnaker) to manage services and software deployments. With the increased adoption of process mining techniques and technologies, in 2019 we will also see more teams experimenting with how process mining discovery, compliance, and performance enhancement can help DevOps teams to learn and improve their CI/CD workflows. — Miguel Valdes Faura, CEO and co-founder, and Charles Souillard, CTO, COO, and co-founder, Bonitasoft
DevOps will play a pivotal role as the enabler of seamless security integration: With approximately 30% of all breaches resulting from a vulnerability at the application layer, organizations will need to adopt a mature, secure software development process that goes beyond just scanning and fixing security flaws. The best way to truly secure applications is to take a programmatic approach to application security that starts with integrating security as early as possible into the software development lifecycle, and to ensure that security and development teams are working hand in hand to make this possible. DevOps will play a pivotal role as the enabler of seamless security integration, and will empower development teams to deliver secure code faster than ever before. Security integration at the beginning of the software development lifecycle will support the overall testing phase and developer workflow. This will ultimately save time and increase velocity in comparison to security testing done at the end of a development lifecycle. As a result, developers will be able to find vulnerabilities during the coding phase instead of during a separate security-hardening sprint, while at the same time ensuring the security of the software being developed, tested and shipped. — Mark Curphey, vice president of strategy, CA Veracode
Adding jobs-as-code to the software delivery lifecycle will avoid the Dev-to-Ops speed-bump: DevOps has proven to be a critical element in helping today’s businesses digitally transform. But one area where DevOps teams still struggle is the Dev-to-Ops handoff into production. Why? Because it’s still too manual. This is why 2019 will be the year when adding "jobs-as-code” to the front end of the software delivery lifecycle will go mainstream. Taking this simple, powerful approach of coding automation instrumentation along with the business logic and infrastructure-as-code, then running it all through the CI/CD tool chain, will help DevOps teams blow by the manual speed-bump into production. In addition, 2019 will be the year when the field narrows down to a leader and a couple of contenders in each stage of the software development lifecycle. The reason is simple: DevOps teams like choice, and as the tools mature they will know which ones they like. Winners will also get the benefit of the ecosystem rallying behind their cause (who doesn’t like a winner?). And finally, enterprise architects will get some well-deserved street cred when they say “governance matters.” — Tim Eusterman, senior director, solutions marketing, BMC Software
The DevOps of the future will be quite different: There’s no doubt in my mind that DevOps is at a crossroads right now, and the DevOps of the future will look quite a bit different from what it was in the past. I think containers are going to become the atomic unit of application execution—the expected standard by both Dev and Ops—and those that are not container-native, for instance a monolithic app running in a virtual machine, will require specialized teams and experience. Adding to this will be serverless functions and these will coexist and be mingled with the containerized microservices creating an increasingly powerful and flexible application architecture, but one which is also more difficult to build and manage. These shifts will necessitate a change in DevOps tools and workflows. As cloud- and container-native applications become the norm, I think the toolchain will naturally follow. Developers will no longer want or need to either install tools or write code on their local machines. Web-based, polyglot IDEs delivered through a SaaS experience (even if that is facilitated by a private enterprise cloud) will be the new default. — Brad Micklea, senior director and lead, Developer Business Unit, Red Hat
2019 will see businesses turning to compliant database DevOps, thanks to two factors: First, the 2018 Accelerate State of DevOps Report from DORA included database development for the first time as a key technical practice which can drive high performance in DevOps. The report revealed that teams which do continuous delivery well use version control for database changes and manage them in the same way as changes to the application. It also showed that integrating database development into software delivery positively contributes to performance, with changes to the database no longer slowing processes down or causing problems during deployments Second, growing concerns about the number and size of data breaches are prompting the introduction of stricter data protection legislation across the globe. Fortunately, the audit trail of changes and automation that DevOps introduces can help to protect personally identifiable information throughout the database development process, provided steps are taken to ensure compliance. — Simon Galbraith, CEO of Redgate Software
Containers will make an impact on CI/CD: The world of CI/CD is being shaken up by container and container orchestration technologies. In 2019, configuration management systems such as Chef, Puppet, and Ansible will be largely replaced by container orchestration tools such as Docker, Kubernetes, and Helm Charts. Docker offers many advantages for applications, while Kubernetes is essential to the successful management of large container deployments. If done right, container orchestration tools can simplify many of the complexities of working with infrastructure. DevOps engineers will have to adapt to this new toolchain. — Jawahar Malhotra, senior vice president of engineering, HackerRank
Container orchestration software will replace many of the functions of DevOps: This year, we saw the adoption of Kubernetes take off as companies of all sizes embraced containers to run cloud-native applications. In 2019, we will start to see container orchestration software replacing many of the traditional functions of DevOps. This will cause a dramatic shift in the world of DevOps as it makes the boundaries between development and operations more clear cut. As the industry moves toward standardized frameworks for software management, it minimizes the work needed to configure and set up software frameworks, leaving more time for DevOps professionals to focus on driving efficient innovation, while also tackling the challenges of managing large, complex clusters of applications across technology stacks. — Dave Smith, VP of engineering, DigitalOcean
This article is the fifth installment of a six-part series by editors of IOUG SELECT and DBTA on "DevOps and the Modern Enterprise."
To access article 1, go here.
To access article 2, go here.
To access article 3, go here.
To access article 4, go here.