Tag: Trace3

  • Unnatural DevOps Delivers Supernatural Results

    Unnatural DevOps Delivers Supernatural Results

    DevOps is not necessarily a “natural” process. Asking people to function outside of their comfort zones and prescribed roles is a challenging task for any organization, and requires a shift in organizational behavior. At Trace3, we believe that while DevOps is not natural, a supernatural result is achievable with the right cultural transformation supported by good tools that drive collaboration and automated flow between people and tools in the application delivery workflow.

    We define DevOps as the application of the principles of continuous flow, feedback and improvement to people, process and technology for increasing agility, stability, security, efficiency, quality and satisfaction. Or, to quote from Michael Schmidt’s article, “Ten ways DevOps will benefit your IT department“: “When implemented properly, DevOps can transform your software engineering process and create value for both employees and customers, producing business performance and value.” These are huge “supernatural” benefits that every business would like to achieve.

    Implementing DevOps properly is challenging. DevOps depends on collaboration; smooth process flow from conception through deployment; and feedback between people, processes and technologies. It’s more natural for people and departments to hoard information and centralize control—traits that are contrary to DevOps best practices. People tend to trust their own abilities and are skeptical when they need to trust dependencies on others. People view competition with others as a means to personal advancement. Compounding this, organizations often manage, measure and reward success at the silo level instead of recognizing a successful delivery as a result of successful execution across the end-to-end DevOps value stream.

    Breaking down silos through transparency and the free exchange of information helps organizations implement the unnatural DevOps processes needed to achieve supernatural results. Fortunately, the transformation is assisted with technology. For example, continuous automation tools liberate sharing of infrastructure, application and compliance information. Pauly Comtois, Hearst’s vice president of DevOps, describes the role of strong tooling in successful DevOps transformations in his article, “Enterprise DevOps at Hearst Digital Media“: “Tools can be an incredible asset in this regard. We deploy Chef across multiple teams in the value stream, automating the deployment, configuration and state of environments in Dev, QA and Production.”

    In scenarios such as the example above, tools support integrated and automated processes for builds, configuration, infrastructure and deployment management. Information managed by continuous automation tools is available for DevOps uses by providing a single “public yet controlled” source of truth for configuration data useful by all processes and tools in the toolchain across the end-to-end pipeline. Those who were concerned and hoarded configuration data can trust that the tools will keep the information secure, reliable and available on demand as readily accessible code.

    Moving beyond the natural behavior of organizations and individuals is difficult, but success can create supernatural results for long-term growth and team effectiveness.

    — Marc Hornbeek

  • Total DevOps is Needed, Plugins are Not Enough

    Total DevOps is Needed, Plugins are Not Enough

    Trace3 was proud to be a Gold Sponsor, exhibitor, trainer of the DevOps Institute leadership certification course and newest partner of CloudBees at the CloudBees Jenkins World 2017 event in San Francisco last month. Jenkins World 2017 attracted 54 sponsors/exhibitors, an 80 percent increase over 2016.

    The growth of interest and sponsors for this event demonstrates that the number of available tool choices for creating the continuous tool chains to realize continuous delivery and DevOps is growing. One of our top takeaways from Jenkins World 2017 is: Plugins are not enough!

    Indeed XebiaLabs’ “Periodic Table of Tools” is a valiant attempt to illustrate categories of DevOps tools and attributes of specific tools. Version 2 of the Periodic Table is now released, and I heard that version 3 is already on the way. While the table shows an impressive 15 categories of tools, I estimate there are actually 29 categories, if you consider tools for cloud management, test management, test creation, code analysis, code collaboration, ALM systems, dashboards, security and others.

    XebiaLabs’ table shows five attributes for each tool: open source, free, freemium, paid and enterprise. In my opinion, there are at least two additional categories: ecosystem and cloud readiness. While the table indicates 120 tools, my own estimate counts 300+ DevOps tools, including CloudBees, Electric Cloud, Perforce, Scalyr, Service Now, Tricentis. SmartBear, SonarQube, Parasoft, Coverity, Fugue, QA Symphony and many others.

    While the exact number of tool categories, tools and attributes is arguable, there are many tools to choose for a continuous delivery toolchain. The state of practice for integrating tools into a seamless toolchain is the plugin or do-it-yourself scripting. Anyone who has implemented a toolchain knows that the existence of tool plugins, while useful, is not enough. The quality and completeness of each plugin varies significantly. Some off-the-shelf plugins are so basic they are little more than marketecture. Even the best plugins are limited to a pairwise matchup between tools of specific version pairs, and not a universal tool or version. For example, a tool with a plugin for Jenkins version X may not function with Jenkins version Y or CloudBees Enterprise of other tools at any version.

    The DevOps Express collaborative initiative announced by 14 participating vendors at Jenkins World 2016 was an attempt to resolve the plugin problem. One year later, the DevOps Express website has shown a growth to 123 tools across 18 categories. However, just like the XebiaLabs Periodic Table of Tools, it is remarkable that the number of tool categories and tools that are not comprehensive (and lacking plugin compatibility) are not guaranteed for any combination of tools that may be needed for a toolchain for a specific enterprise.

    Each enterprise has unique toolchain requirements and preferences. Selecting and an integrating DevOps tools is not like buying a box of “standard” Lego bricks that are guaranteed by the manufacturer to fit together. Each tool and plugin is different.

    The choice of tools, the combination of tools and the integration of the chosen tools all must be carefully chosen to fit the specific customer requirement. More importantly, the integration must be architected to ensure the chosen combination works well together in a seamless toolchain.

    Until there is an absolute standard plugin architecture that all tool plugins are guaranteed to be compliant with, expert consulting services will be essential to make tools choices and make sure the tool chain works.

    With such a rapidly growing list of tools and vendors, only the most capable consulting teams with deep insight and partnerships with the widest array of tool vendors will be able to address enterprise DevOps solutions properly.

    — Marc Hornbeek