2020 DevOps Predictions
At XebiaLabs, we not only like to reflect on years past, but we love to keep our fingers on the pulse of the future! Here's what our in-house experts predict about what 2020 has in store for the DevOps industry.
The DevOps as a Service model will continue to be the North Star for our customers. Small 2-5-person teams will serve up DevOps tools and then everyone else will consume them. This is how a software delivery process remains or becomes compliant and secure. No longer will each team inside an organization be doing its own thing. DevOps as a Service provides a centralized model to the entire business.
Data collection and analysis is going to become even more important in orchestrating CI/CD pipelines and the delivery process as a whole. There's always more data, but in 2020, it will be about what to actually do with it and what will change in individual organizations and the industry. Currently, the majority of data collected is from the right. There will be a shift left in data collection to implement testing, change, and compliance earlier in the delivery process.
Also, the lizard people are going to take over the world.
Giving developers freedom while maintaining control will be crucial. You can have both! Development can have the freedom to leverage the tools they like and the platforms they want. At the same time, from my perspective, XebiaLabs can put a layer on top that gives leadership in the organization the confidence that they're getting what they need. Balance can be struck. Big organizations have invested heavily in tools like Jenkins and Jira for years and years - they've become fundamental to how they run. These companies are not willing to change out these tools, and they shouldn't have to. We integrate with what they're already using PLUS offer enterprise features that encourage scalability, compliance, and speed.
Development is so critical to the success of the organization that they are a part of because all companies have become software companies. Without engineers, there's no business, so they need to be productive and and that requires working the way they want to. However, currently, there's not a lot of visibility into the development pipeline. With the XebiaLabs DevOps Platform, we add that visibility without disrupting development teams. The business gets what they want, and Dev gets what they want. You don't have to make a major change to quickly bring efficiency to your software delivery process.
I see three big trends ahead.
- Data. DevOps and DataOps will converge! There's so much data out there across all technologies and spaces, and we need to have better ways of managing it, collecting it, searching through it, and making it visible to others. The next wave of DevOps will involve who can get this right. Who can enable large enterprises to take full advantage of their data? Well, we think we've got it with our Software Chain of Custody capabilities.
- Chaos engineering. Heading into 2020, chaos engineering will be the next wave in security. Building failures into your production system shows that you are confident that your system can overcome them! I predict that chaos engineering won't become mainstream yet, but that it will become a more regular part of the delivery pipeline within startups, and tested through the first stages of development in larger companies.
- Culture over tech. This new year will bring an even larger focus on culture over technology in DevOps. Of Gene Kim's 5 Ideals in his new book, The Unicorn Project, two of them focus on culture. For a robust DevOps journey, your people must not be afraid of failure and what comes with it - it's all part of the work. Accepting and fixing failures without fear can only produce more quality work moving forward. Technology has brought DevOps practices to the mainstream, and moving forward, we need a similar effort to create cultural change that really uplevels` DevOps initiatives.
Embracing accelerated teams will become mainstream
Speed has mattered for years. Acknowledging developer speed as a crucial parameter to accelerate software delivery as a whole is becoming mainstream. In many organizations, islands of speed already exist. Organizations can benefit by looking into the three types of accelerated teams and embracing their practices:
- The Innovators. These are the teams that push modern microservices applications to the cloud. They have embraced modern CI tools and have the ability to release whenever needed. The simplification of the application landscape and the ability to release independently are two ingredients that can be applied to almost all application stacks. Principles like GitOps and NoOps are two vital ingredients that they live by day and night.
- The Automators. These are the teams that take automation extremely seriously and have automated the full end-to-end pipeline. They’ve included GUI testing, acceptance testing, linked systems (like ServiceNow) together, and have shown how you can still automated full end-to-end cycles in an existing enterprise. Even with full automation and connection, there are still some wait times and bottlenecks due to dependencies on other disciplines next to Development.
- The Simplifiers. These teams challenge the status quo. They don’t blindly follow rules and guidelines that have existed for ten years as a given – they challenge them. They are motivated not to escape from the rules, but to simplify the rules. Embrace the challenge! Reconsider and revise what you defined before.
All three teams combined are gold for your organization. A simplified, automated process will allow you to embrace innovation and accelerate as a whole.
Growth in understanding, managing, and reducing complexity will become crucial
A primary roadblock for acceleration of software delivery is complexity. This complexity appears in three layers:
- Application build dependencies are at the heart of complexity and make for complex releases.
- Business ownership and business application stacks drive end-to-end testing and approval dependencies that slow you down.
- Shared environment and infrastructure resources.
Moving forward in 2020, understanding the complexity is the first step. Using systems to actively build an understanding of the release cycle, the software delivery as a whole, and deployment patterns is vital to start managing it.
Next comes simplifying the landscape. Get rid of central environments and big release moments. Strive for Single Artifact Push to Production – this is an ultimate goal many teams can set. It is vital to invest time into composing solid pipelines that automate everything.
- Register for our webinar DevOps 2020
- Check out our 2019 DevOps Reflections
- Read the white paper: The IT Manager’s Guide to DevOps