Author: Eric Maxwell

  • Driving Organizational Success With the 2022 State of DevOps Report

    Driving Organizational Success With the 2022 State of DevOps Report

    “Ship it!”

    We have just completed the research and analysis for the 2022 Accelerate State of DevOps Report (SODR) and oh boy, do we have some interesting things to share! First, let me take a short step back for those of you that may be new to the DevOps Research and Assessment (DORA) and our annual report, the SODR. DORA is an academically rigorous research program that seeks to answer the questions, “What practices enable teams to be high-performing software teams? And how do these practices impact organizational performance?”

    Since 2014, we have surveyed 33,000 practitioners around the world spanning all major industries. The project is intentionally–and staunchly–tool- and platform-agnostic. The research builds on itself, and each year we seek to both evaluate previous findings and expand in other directions or areas of research. This is important to note because many of the core findings, such as “software delivery performance drives organizational performance,” have been validated year after year. I would encourage you to go back and read all the reports from years past.

    Security

    This year, we doubled down on our 2021 research into software supply chain security by looking at the technical practices that improve software supply chain security and the non-technical practices that impact an organization’s ability to excel at securing their software supply chains. We leveraged two frameworks to focus our research: Supply Chain Levels for Software Artifacts (SLSA) and NIST’s Secure Software Development Framework (SSDF). Below is a summary of some of our key findings in this area:

    • Shifting left on security is a broadly adopted practice. Our research shows that two-thirds of respondents are actively pursuing software supply chain security by seamlessly baking security into the development process.
    • Culture is the primary driver of security practice adoption. One might expect that technology is the primary driver, but our research showed that a generative organizational culture (eg. performance-oriented, highly cooperative, shared risk) leads to healthier software practices.
    • Technical practices around CI/CD predict success in security. Companies that use source control, continuous integration (CI) and continuous delivery (CD) have more established SLSA practices. These practices shift security left to the developers and ensure consistent security scanning.
    • Cloud enables secure software practices. The five characteristics of cloud computing as defined by NIST enable the successful adoption of software supply chain security which, in turn, predicts greater organizational performance.

    Our research showed that companies who prioritize and excel at securing the software supply chain experienced fewer service outages, anticipate fewer security breaches and demonstrate high levels of both software delivery performance and organizational performance. The data showed that through the use of modern practices like continuous integration, teams can improve their security posture and even amplify the positive impact these security practices have on software delivery metrics (MTTR, deployment frequency, lead time-to-restore service) and overall organizational performance.

    Culture

    We saw above that the biggest predictor of an organization’s application development security practices was the presence of a generative organizational culture based on shared risk and information sharing. We also see that elements of these types of cultures lead to higher overall organizational performance, as well. Our research showed that high organizational performance can be achieved by fostering environments that are:

    • Supportive: Teams that felt supported and that they had buy-in from leadership (e.g., more financial support, more allocation of resources, sponsorships, etc.) were associated with high-performing organizations.
    • Stable: Teams that did not experience much change in membership over the last 12 months were more likely to be a part of high-performing organizations.
    • Flexible: Organizations with higher levels of working flexibility around where work was performed–remote, in-person or hybrid–were higher-performing overall.

    We also looked at burnout again this year and expanded the scope to understand which elements of culture contributed to lower levels of burnout. We found that generative culture, team stability and work flexibility all contributed to a reduction in burnout among employees.

    Reliability

    In previous years, our research told us that those that excelled in technical practices also excelled at organizational performance. This year, we have more nuanced data about this topic. This year, we saw that software delivery does not predict strong business outcomes unless these practices were paired with reliability. Think about it–will a customer be satisfied with new features if the service isn’t stable? What is the benefit of pushing code quickly into a fragile environment? Reliability is an essential component of driving organizational performance through software delivery performance.

    We also see that site reliability engineering’s (SRE’s) impact on organizational performance is non-linear; reliability engineering practices often do not result in additional reliability or organizational performance until a certain maturity is reached. It’s important for teams to know this and approach their SRE practice as an investment. It will likely not be sparkly ponies and unicorns initially as you build the reliability muscle but, as you advance, high performance and success are likely.

    Cloud

    Public cloud use is up a whopping 36% over 2021, whereas companies reporting no cloud usage at all are down by 50%. The use of hybrid cloud is up by 25%. Unsurprisingly, the use of cloud computing was associated with higher organizational performance. In past years we saw that it wasn’t “using cloud” per se that led to organizational performance, rather it was the achievement of the five essential characteristics of cloud computing–on-demand self-service, broad network access, resource pooling, rapid elasticity and measured service. This year, we saw that cloud computing enabled things like reliability, continuous delivery and improved supply chain security which drive organizational success.

    Over 50% of folks who took our survey reported that they leveraged multiple cloud providers. We asked respondents what benefits they realized from using multiple cloud providers. Here are the top three:

    1. Availability
    2. Leveraging the unique benefits of each provider
    3. Trust is spread across multiple providers

    Given the fact that reliability seemed to be the key to being a successful software delivery shop as well as a successful organization, it is not surprising to see availability listed by almost 63% of respondents as a benefit derived from using multiple clouds.

    Summary

    Our research continues to progress and we continue to dig deeper and deeper into the capabilities and practices that impact your business. We see broad adoption on the topics of security and reliability which are both heavily anchored in culture. Good culture leads to success. Period. There is no magic culture button and it takes work, but it is very achievable. Start by defining organization-wide or line-of-business-wide goals and then just start continuously improving. Don’t worry about creating a three-year plan for improvement; create a one-month plan and just get to work instead. At the end of that month, evaluate your learnings, refine your focus and get back to work. With a commitment to hard work and continuous improvement, your investments should start paying off and, as an added bonus, you’ll have a better culture.

    We hope you have enjoyed these juicy findings as much as we have and we encourage you to read the entire 2022 State of DevOps Report.

    Also, please join us at dora.community to continue the discussion about these findings and to share and discuss your experiences on your journey toward excellent software delivery and operations.

  • Accelerate State of DevOps Report: The Move to Multi-Cloud

    Accelerate State of DevOps Report: The Move to Multi-Cloud

    Analysis of the 2021 Accelerate State of DevOps Report (SODR) data shows an increasing number of organizations are choosing multi-cloud and hybrid cloud solutions. In our survey, respondents were asked where their primary service or application was hosted, and the results showed that public cloud use is on the rise.

    Of the 1,200 survey respondents this year, 55% indicated they’re using a public cloud, a 5% increase from 2019. Of those approximately 660 people, 21% indicate they use multiple cloud providers and 34% indicate using a hybrid cloud strategy. For clarification, multi-cloud refers to the use of two or more public cloud providers and hybrid cloud refers to the use of an on-premises data center and/or private cloud in combination with one or more public clouds.

    multi-cloud adoption

    Our data indicated that the No. 1 reason for using multiple providers is to leverage the unique benefits that each provider brings to the market. Every cloud provider provides basic compute and storage services, but building a best-in-class business can be best achieved by taking advantage of best-in-class cloud computing features and not being locked into a specific vendor. In fact, 26% of our respondents indicated this was their primary reason for their multi-cloud or hybrid cloud approach.multi-cloud reasons

    The second most-reported reason, at 22%, was availability. Companies around the world are increasingly identifying as software companies and relying on software to be their primary business driver. Software allows companies to deliver value to their customers, outperform their competition, respond to market events and bring innovation to the market. Availability of this core product is not a “nice to have”, it is essential. The truth is, however, that computers and networks are fragile—they fail. If we accept this fact, we can design and build systems that are prepared for this eventuality. Companies are addressing these risks with strategies that decrease the blast radius of changes and allow systems to continue operating even in the face of failure.

    Speaking of worst-case scenarios, disaster recovery (DR) is the third most-reported reason for adopting multiple cloud providers. In fact, the cloud and the global proliferation of data centers has allowed companies to rethink what disaster recovery actually means. Gone are the days of systems going down and operators scrambling to rebuild systems and restore from tape backups—the cloud is the backup. The adoption of cloud plus fully automated pipelines creates a paradigm shift around DR that allows companies to design their disaster recovery strategy around zero downtime.

    In addition to hybrid and multi-cloud strategies increasing, we also saw a rise in the use of the five essential characteristics of cloud computing. The National Institute of Standards and Technology (NIST) defines the five essential characteristics as:

    These five characteristics must be embraced and used to realize the value of the cloud. A company that adopts cloud but does not employ these characteristics simply shifts their operating environment from one place to another but doesn’t gain anything. In fact, the unfamiliarity of a new environment, new pricing models and lack of optimization from years of running on-premises often leads to disappointment. Successful companies embrace these five characteristics.

    We know from previous reports that these characteristics are essential for delivering software at a high or elite performance level. We see adoption rates between 73% and 78% across the five characteristics with the lowest being on-demand self-service (73%) and the highest being measured service (78%). There is no doubt that this increased adoption is a key driver for the jaw-dropping performance we saw this year from our elite performers.

    Elite performers can now deploy 973x times more frequently, have a 6,570x faster lead time to deploy, have a 3x lower change failure rate and a whopping 6,570x faster time to recover from incidents when failure does happen. Go back and read that again—no, those are not typos nor an error in the data! An elite-performing company can bring a new feature to market 6,570x faster than a low-performing company. This is one of the reasons we see a direct link between software delivery performance and overall business performance.

    We encourage you to read the entire 2021 Accelerate State of DevOps report. There’s lots of great information in the report including the addition of a fifth key metric—that’s right; the ‘Four Keys’ are now the ‘Five Keys.’ “What is the new metric?” you ask? Well, I guess you will have to read the report and find out. 🙂