Overcoming DevOps and Software Delivery Challenges

Multiple research reports have been released in recent weeks that identify challenges and opportunities for DevOps, especially in the context of the COVID-19 pandemic.

One key finding from the reports is that COVID-19 has accelerated DevOps' move away from on-premises deployments to the cloud. According to the Codefresh second annual State of DevOps survey released on Aug. 18, due to the pandemic, 58% of DevOps professionals are now planning on moving some infrastructure to the cloud.

The Codefresh survey also found budgets for DevOps efforts are on the rise in 2020, with 74% of organizations reporting that they expect DevOps spending to increase this year. With more money allocated to DevOps efforts, there are a few areas that need help. A key promise of DevOps is automation for development, but 67% of the Codefresh survey respondents reported that they spend 25% or more of their time addressing bugs in automated systems.

Software Delivery Challenges

The 2020 State of Software Delivery Management Report, commissioned by CloudBees, found that one of the challenges organizations face is a lack of information, and that is impacting software development and delivery.

Shawn Ahmed, senior vice president and general manager at CloudBees, told ITPro Today that one surprise in the report was that so many (84%) of the respondents found inaccessibility to critical information limits their ability to make decisions. The lack of visibility also is reflected in the finding that more than 50% of the study's respondents couldn't compare the time developers spent dealing with technical debt and bugs to time spent developing new features.

Adding further insult to injury, the report found that 65% of organizations were unable to quantify the cost to the organization of a feature delay in development.

There are a number of potential reasons why developer visibility is limited, according to Ahmed.

"It's a confluence of things like available analytics expertise, as well as the time and resources it will invariably take to build insights," Ahmed said. "Data is a hard problem to solve. You have to first access it, then map and model it, connect it to the outcomes you want to see, and make sense of it through context to gain the most pertinent insights."

Another path to improving software delivery outcomes, according to the report, is embracing the software delivery management (SDM) model. Sixty-one percent of the survey respondents noted that SDM approaches improved delivery times for software. SDM should not be confused with the concept of software development life cycle (SDLC), according to Ahmed.

The SDLC model is about the process of building and delivering software, he said. SDLC is a framework that defines the tasks performed at each step in that process, while SDM is the principle of managing the process.

The relationship of SDLC to SDM is akin to the process of traditional manufacturing, Ahmed added. There are those that are involved in the building of a “thing” and those that manage the building of that “thing.”

"Now, it’s not that this hasn't been happening all along, but the survey clearly indicates that there is much optimization to gain still, and this is an area where purpose-built SDM software can help teams improve by an order of magnitude," Ahmed said.