Streamlining product development lifecycles with automation

Thousands of people at IBM support the development of IBM® z Systems® and managing the engineering workflow for such a large enterprise is an enormous undertaking. To keep releases on schedule, IBM Z platform teams must constantly coordinate manufacturing, chip design, hardware, firmware, OS, testing, defect tracking and other project threads and process massive amounts of data.

Unrelenting delivery deadlines mean that issues and errors must be addressed early in the development lifecycle. Systems must also meet stringent regulatory requirements across a variety of industry and governmental standards, so testing, validation and verification must be comprehensive.

“There are lots of government controls around export—who can be involved in development restrictions for some countries, export control licensing for others as an example,” says Chris Roberts, Architect, IBM Systems DevOps Enterprise Solutions Systems. “We have to be audit compliant and very secure. IBM is active in the open-source community, but we don’t want people looking at our proprietary chip designs. We have to isolate things, even inside IBM, so that some development teams can’t see what others are doing.”

For years, the zHW platform team relied on several workflow management tools to track development—but each had shortcomings. “One of the solutions that IBM acquired years ago was very powerful and did everything we wanted, but it was very complex and not intuitive,” says Dominic Odescalchi, Project Executive and Manager, IBM zHW Program Management. “It had challenges with usability, so people started to look at other options.”

Working across the IBM Z development enterprise, key stakeholders created an evaluation matrix chart. It presented a side-by-side comparison of workflow tools and showed which solutions satisfied and which did not meet the integration capabilities for the tools used by the team. “This was a year-long effort to make sure we selected the right tool to provide the outcomes we were looking for,” says Odescalchi. “We also had the freedom to choose the best tool that satisfied our must-have requirements.”

After receiving input from team members around the world, a minimum viable product (MVP) was defined that met the key criteria of usability, scalability, vertical and horizontal integration, security, cost, backlog prioritization, command line support, technical support, and data and dependencies management.

Although a broad cross section of IBM z Portfolio teams participated in the creation of the MVP, the project started with the zHW teams, which were committed to transitioning to the new tool for the first pass of implementation.

Based on a year-long global evaluation process, the IBM Engineering Workflow Management (EWM) tool stack was selected as the solution. “By being completely objective, and allowing the criteria and data do the talking, we were led to EWM,” says Odescalchi. “EWM was the consensus tool that we collectively agreed upon to provide the best solution.”

Of all the selection criteria, one of the most important factors was scalability. “We need an enterprise-scale solution and not many tools can support our volume of data and concurrent users while still providing the performance required,” says Odescalchi.

The zHW platform leadership team responded to the issue of complexity by developing personas for key EWM system users. “The focus on personas enabled us to achieve usability and simplicity for people inputting data as well as people consuming data,” says Odescalchi. “They won’t be overwhelmed with dozens of menu options and selections that are not relevant to their persona.”

Another key MVP requirement was that the track-and-plan and defect management systems needed to operate as one integrated environment. “We did not want those to be two disparate systems and EWM provided that singular environment,” says Odescalchi.

Due to the very large file sizes created during testing, any solution would have to support large data storage attachments. “The data has to be readily available in working storage. A lot of the other tools required archiving after only one year, which falls short of the capabilities we need,” says Odescalchi. “IBM EWM can store years’ worth of real-time information without having to archive, and none of the other tools came close to providing that capability.”

IBM EWM software will coordinate the central hub of engineering data for the zHW platform development team and works in tandem with the IBM Engineering Lifecycle Management (ELM) solution. “Some teams might be using an agile process that has a totally different workflow than waterfall, but it still has to plug in and be consumed by the other teams,” says Roberts. “Workflow customization within ELM helps the solution adapt to every team while still coordinating one view of the development data and progress.”

Finally, to ensure that hardware is ready to release, the team completes testing with the IBM Engineering Test Management (ETM) solution. “Because, again, this is hardware. You end up with a physical thing that you’ve developed. As such, things need to be well tested—and everything has to come out right the first time,” says Roberts.