Ramkumar Venkataraman, Senior Vice President and IT Transformation Leader at Moody’s Corporation, discussed how DevOps can be implemented to transform the legacy space.
Venkataraman began his keynote address at the 2016 Chief Information Officer Leadership Forum: Spotlight on Financial Services, held on September 8 in New York, by commenting that whenever he speaks about DevOps at conferences, people are interested in the topic but they always respond with, “We have so much legacy. How can DevOps solve our problems in that space?” Venkataraman had some answers to that question and explained how he implemented DevOps at his own company, working with legacy applications.
“The first thing we had to do was define what we saw as our goal in DevOps. For us, it was improving collaboration,” explained Venkataraman. “We wanted to use DevOps to create the kind of innovation we couldn’t otherwise create. I told my operations team that the objective of this entire program is to eliminate our jobs—or eliminate the way we do those jobs today. We’re going to do things much smarter and get rid of the repetitive, mundane things we do day to day,” he said.
“I told my operations team that the objective of this entire program is to eliminate our jobs—or eliminate the way we do those jobs today.”
“DevOps isn’t about toolsets,” stated Venkataraman. “It’s about how we do things—finding opportunities and overcoming obstacles by bringing people together to solve problems. Implementing DevOps concepts can transform legacy to provide more services for business value enhancement.” Venkataraman pointed out there’s a huge opportunity to reduce lights-on and free up bandwidth for other business functions.
There are many challenges to applying DevOps to legacy applications, including long lead time to delivery, error-prone manual processes, high dependence on people (small tasks require a specialist), and lights-on costs that reduce business value delivery. “We’ve initiated a fundamental shift in how IT provides services to meet business objectives. Our DevOps program is cosponsored by development and infrastructure teams with a cross-functional steering committee cutting across all groups. We then set up our DevOps scope. It used to take us four to six months to set up our test environment (with no automation); it now takes us a month and a half,” stated Venkataraman. He summarized the following successes of his company’s DevOps implementation program within the first six months:
• Release cycle decreased from more than 25 hours to less than eight hours, in the legacy space; in the greenfield space, deployment is one to two hours
• Data load optimization—a 30% reduction in load times
• Test environment availability is 98%, up from 70%, as a result of automation
• Integrated platforms in four weeks rather than 25 weeks
• Eliminated manual effort on high-risk tasks
• Reliable management of certificate expiration
• Automated stop and start of servers
• Minimized need for business testing post-deployment
• More resources available for business value addition
• Reduced by 80% change-caused incidents
• 12% operations cost reduction for all legacy applications; next year’s goal is 30%
• Access provisioning automated, reducing the process from two days to five minutes
• Dramatically reduced outage during patching
“We’ve initiated a fundamental shift in how IT provides services to meet business objectives.”
“The primary key to successfully implementing DevOps concepts in the legacy space begins with a change in mindset,” Venkataraman stressed. Other keys to success include collaboration and innovation to solve problems, celebrating small wins to maintain enthusiasm, self-funding, quick decision making, and education/training. “Quick decision making is at the crux of all of this,” said Venkataraman. “This is DevOps’ contribution to the whole process.”
“The primary key to successfully implementing DevOps concepts in the legacy space begins with a change in mindset.”
In summary, Venkataraman delineated his company’s steps for the future to achieve 30% to 35% total reduction of lights-on costs while improving stability, availability, and speed of delivery:
• Self-healing of the systems (detection to restoration without human intervention) to minimize outages
• Self-service of the systems (deployments and access without operations teams)
• Provisioning platforms in two weeks
• Auto configuration set-up, monitoring, and management
• Zero inconsistencies among test environments
“Lastly, we’ve invited every employee to be part of the journey.”
Visit Argyle Executive Forum's Argyle Digital: CIO Leadership Forum in Virtual Forum, on Jul 07, 2020