Executive Insights on Scaling DevOps
Automation, culture, collaboration, and tools are the most important elements of DevOps.
Join the DZone community and get the full member experience.Join For Free
To gather insights on the current and future state of how DevOps is scaling in the enterprise, I asked IT executives from 32 different companies to share their thoughts. Here’s who shared their insights:
- Nancy Wang, CEO, Advancing Women In Product (AWIP)
- Mick Morrissey, Director of Engineering, Asavie
- Lyon Wong, Founder & COO, Blameless
- Patrick Reister, Senior Build EngineerandIvan Szatmári, Head of Release Management, Bohemia Interactive Simulations (BISim)
- Rob Zuber, CTO, CircleCI
- Brian Dawson, Director of Product Marketing and Brian Nash, Director of Product Marketing, CloudBees
- Eric Robertson, V.P. Product Marketing Management & Strategy Execution, CollabNet VersionOne
- Jeff Williams, Co-founder & CTO, Contrast Security
- Mike Rose, VP Engineering, Cybera
- OJ Ngo, CTO, DH2i
- Chris DeRamus, Co-founder & CTO, DivvyCloud
- Tobi Knaup,Co-Founder & CTO, D2iQ
- Andi Grabner, DevOps Activist, Dynatrace
- Antony Edwards, COO, Eggplant
- KrisLahiri, Co-Founder & VP Operations and Chief Security Officer,Egnyte
- Chris Michael, DevOps Engineer at FileCloud
- Tamas Cser, the Founder & CEO ofFunctionize
- Justin Stone, Senior Director of Secure DevOps Platforms, Liberty Mutual
- Mark Levy, Director of Strategy, Software Delivery, Micro Focus
- Phaedra Divras, Chief Operating Officer,Mission
- Michael Morris, Senior Director of IT Cloud & DevOps Platforms, NetApp
- Tori Wieldt, Developer Advocate & Sr. Solutions Marketing Manager, New Relic
- Bob Davis, CMO, Plutora
- Veejay Jadhaw, CTO, Provenir
- Vishnu Nallani Chekravarthula, V.P. Head of Innovation, Qentelli
- Anurag Goel, Founder & CEO, Render
- Davy Hua, Director of DevOps, Security & ITOps, ShiftLeft
- Dave Karow, CD Evangelist, Split.io
- Ben Newton, Director of Product Marketing,Sumo Logic
- Adityashankar Kini, V.P. Engineering, Sysdig
- Neil Barton, CTO, WhereScape
- Dan Beauregard, DevOps Evangelist, Xebia Labs
You may also enjoy: A Beginner's Guide to Scaling DevOps
Here’s what we learned:
- A litany of things, including automation, culture, collaboration, and tools are most frequently mentioned as the important elements of DevOps. Automation, cloud, and culture all lead to speed. Culture, lean, automation, measurement, and sharing all create an environment for innovation. DevOps is the understanding that with automation in place, operations can be done more quickly and more predictably without human error. Without a team that buys into the importance of culture where collaboration is paramount, there is little point to discuss CI/CD tools. DevOps is a cultural thing. If you go into a cross-functional cultural conversation with different goals, it will not work. Make sure you are building a culture of collaboration across all functions.
- Likewise, automation and culture get more mature and sophisticated as the DevOps methodology scales within an organization. Over time, people see the benefits of DevOps and stop fighting it. The culture improves as the DevOps technology matures and people become better educated and less terrified. There’s less fear of failure and greater anticipation of success. The longer you go on with DevOps, the more integrated it becomes. We moved from separate product and DevOps teams to more integrated where everything is in the Scrum team, there is not a dedicated DevOps team. Having developers in the same room with product management and customer success has buoyed the kind of culture required to make DevOps successful.
- More integrated toolsets, automation, and data are needed to successfully scale DevOps. Tools that tie toolchains together, take data from different tools and apply metrics to reduce waste, improve flow, and deliver value are immeasurable. Once you start gaining maturity, you start seeing more home-grown tools and more standardization across tools. Cross-team collaboration and agreement to use the same tools and processes matter much more than which specific tools are used. Use tools to automate as many processes and procedures as possible. Decouple deployment from release, build in feedback loops, measure everything. Data transparency is key for monitoring feedback loops. Measuring is the first step towards continuous improvement and often requires automation, both to ensure the process you’re measuring is consistently done is exactly the same way but also to make relevant metrics accessible. Data must be easy to access and see along with the operations and delivery pipeline.
- The most frequently mentioned use cases for a scaled DevOps methodology is reduced mean-time-to-resolution and security while the most frequently mentioned industries where scaled DevOps is having an impact is insurance, IT, and entertainment. Meantime, to repair or resolve can be reduced by 80 to 95% with monitoring the production environment and automating the remediation. Automation accelerates the learning process which reduces MTTR. When a team has greater insights into their entire stack, they are able to troubleshoot more effectively and take more precautions to ensure they don’t have outages, and if they do, they can recover more rapidly. We help customers to quickly and efficiently deploy apps from development to deployment to production and if there are any changes or challenges, they can quickly go back. This results in less malware and security risk. Shifting security left integrates cloud security checks during the CI/CD process before code and infrastructure hits runtime.
- Consistent with getting started with DevOps, the most common issue when scaling DevOps continues to be culture and adoption along with skills, tools, coordination, and automation. Cultural and organizational issues are ongoing even as DevOps becomes more ingrained. Teams need to focus on ownership, education, access, and data to further enhance their ability to help build and support their solutions. People are afraid of losing their jobs if you incorporate automation. People need coding skills and need to reinvest in education. It takes time and money to see the results at scale. Take the time to choose the right tools for the job. Bring the entire company under the same architecture, building practices, and application landscape. This takes a lot of time and effort. You have to use automation, especially for less frequent tasks that can be associated with higher risk or if the person who did it before is no longer on board.
- Concerns around scaling DevOps are around its difficulty, security, vague definition and understanding, and its relation to business problems. DevOps is still hard. Even when you get past people’s unwillingness to change, they have to collaborate, use different tools, incorporate security and QA earlier in addition to looking at containers, cloud, and serverless architecture in the cloud. This is especially challenging for traditional enterprise IT organizations. Security elements need to be fully integrated and automated into the DevOps pipeline. More teams should have security professionals to implement DevSecOps. There’s a danger DevOps is becoming so vague it becomes meaningless. Ensure everyone has the same understanding and definition of DevOps. If different parts of the organization have a different definition or understanding, you are doomed to fail. We need to focus on the business of DevOps, what it costs, how it’s measured, the business problems being solved. The organization has to change how they approach the challenges and integrate DevOps in the initial design stages.
- The future of DevOps is more automation, ingrained security, and shorter feedback loops with data being used to drive automation. Except for the culture piece, everything will be automated similar to distributed automated organizations. You want ore automation around deploying and installing software and the ability to take action. Security professionals will be integrated into DevOps teams and these teams will learn identity, authorization, and secrets management, as well as overarching risk and compliance to ensure the individual and the teams have the skills, access, ownership, and data to make the right decisions. Telemetry will be hooked up to everything collecting data to make more informed decisions. We’ll have power tools for feedback loops. We’ll know the future has arrived when we don’t need the term “DevOps” anymore. The process and tools which we currently lump together under the term DevOps will simply be employed on a day-to-day basis without anyone making a big fuss about it.
- Developers need to keep learning, embrace automation, understand the business, and think about end-to-end integration of the pipeline. Educate yourself on best practices. Ask engineers to expose you to what happens post-deploy. Educate yourself on how modern deployment models work – rolling versus blue jeans. Listen before you give your opinion. Care about both business value and usability. Reduce the need for human intervention. Automate testing. Understand how cloud testing can be used. Understand the business and what’s critical to its success. Take responsibility for the entire pipeline and value chain of software development and product engineering. When everyone who is building a product cares about both aspects of quality, cares about what happens when it gets deployed, and gets exposed to the end-user is the result of true DevOps adoption.
Opinions expressed by DZone contributors are their own.