everythingpossible - Fotolia

Evaluate Weigh the pros and cons of technologies, products and projects you are considering.

How value stream analysis unclogs DevOps workflows

Break through cloud DevOps workflow bottlenecks by following the value stream, not the yellow brick road. Find out how value stream analysis undams app development.

Enterprise architects often focus their DevOps efforts on improving the cloud application development processes that are highly visible and easiest to automate. Practices like continuous integration and automated testing can dramatically improve certain components of the DevOps workflow. However, without visibility into all the processes between business ideation and deployment of new functionality, app development workflows can get stuck. Using cloud-based value stream analysis processes can uncover bottlenecks and keep development projects on track.

Many enterprises are starting to apply value stream analysis to cloud and on-premises app development processes to better manage DevOps workflows. Along with expert advice on this process, this article looks at a company, Barclays PLC, that is doing it. All those quoted were speakers at the recent DevOps Enterprise Summit 2017 in San Francisco.

A value stream is, at its essence, a set of processes for creating and producing a product or service to market. Key steps in value stream analysis include creating flowcharts and maps. Step one, the flowchart, calls for diagramming the value-added and nonvalue processes in a business's current app development workflow. Value stream mapping adds greater details about the current state of the DevOps workflow and goals for a future, waste-free app and microservices development workflow.

Focus on value-tivity

Process improvement is always important, but creating business value should be a higher priority in DevOps workflows, according to Jonathan Smart, head of development services at Barclays PLC. That's where value stream analysis plays well.

"It's not about productivity; it's about value-tivity," Smart said. "Productivity represents the capacity to increase output for a given level of input. But this output can also be rubbish, which adds no value to the business."

While some project management teams get so efficient that they front-run other business units or partners and do work quickly, that does not add business value. Barclays discovered that new features got delayed by a security and auditing process at the end, Smart said. When problems were discovered, the teams had to do more work to fix them.

Smart advised that DevOps leaders embed security and governance experts into development teams from the get-go. These experts can analyze a specific project at the beginning and better identify security and governance requirements. Previously, when security and governance experts were managed separately from development at Barclays, it was much harder to plan the release of new features. Those security and governance experts shifted between mortgage products one week and financial trading the next. With embedded experts on the value stream analysis team, Barclays allows developers to focus their attention on the problems that matter most.

Going with the value flow

A focus on the IT value stream also allows Barclays to quickly adapt to new business ideas, Smart said. Previously, a new team member would be assigned at the tail of an update and would have to familiarize themselves with all the relevant rules. "There was no appreciation of the value stream. Now, someone from each function works together in a long-lived team associated with long-lived products," Smart said.

This way of looking at all the steps associated with different kinds of functionality also makes it easier to spin up a new value stream required to quickly execute new business value, according to Smart. Now, at Barclays, when there is a market disruption or new regulation, the organization can shape shift, break value streams apart and create new ones.

Using visualization in DevOps workflows

Value stream analysis is much harder when you have thousands of developers working on hundreds of different products.
Mik KerstenCEO, Tasktop

Getting a complete visualization of an app development workflow is not easy, said Julia Wester, a consultant at lean management provider LeanKit. Leaving anything out of visualizations leads to a failure to analyze the impact of change. One good practice, she noted, lies in leveraging tools that help automate the tracking process across different cloud management tools. This can help leverage a more complete picture of what the organization is doing.

Automating tracking of app development workflows allows managers to get a better sense of when processes are out of alignment, according to Wester. She recalled visiting a customer whose development and operations teams were moving so fast that they were out of alignment. "It was a case of your Agile is bumping up against my Agile," she said.

In value stream analysis, Wester said, managers need to investigate the handoff point between processes. "If we are only looking at one or two touch points of a process, we will not make the massive improvements we could have if we really understood and optimized all of the touch points in our systems."

Tools, staff can clog value streams

In startups, it is much easier to do value stream analysis because the DevOps workflow is simple, according to Mik Kersten, CEO of Tasktop, an integration hub provider. "Value stream analysis is much harder when you have thousands of developers working on hundreds of different products."

In his work with large companies' app development teams, Kersten has seen many organizations have an ad hoc approach to identifying, analyzing and managing the value stream. That's because the complete systems development lifecycle can include so many tools -- tools for project management, portfolio management, requirements analysis, multiple flavors of Agile, quality, release management, deployment management, continuous integration/continuous delivery, application performance management and more.

Because there are so many tools involved in DevOps processes, "the value stream can get buried," Kersten said. Even worse, among those tools can be some tool that can hinder a smooth DevOps workflow. For example, he noted: "Jira is a great develop[ment] tool but [is] horrendously bad at requirements management."

Don't forget the human element that can contribute to DevOps workflow slowdowns, the experts said. DevOps teams include many specialists, experts in business, customers, designers, developers and operations personnel. Be sure to include a value stream owner in each project's team to make sure that the team is focusing on the right business values.

Dig Deeper on Managed Kubernetes and container services