The McKinsey paper, Asia-Pacific Banking Review 2019 Bracing for consolidation: The quest for scale, says banks need to improve the way they develop and deliver solutions by implementing the latest DevOps tools and methods across the development pipeline.
It suggests that improvements can mean up to 80% improvement in time-to-market, reduction of up to 75% in time spent diagnosing systems and cutting FTE (Full Time Employment) by 75%.
One technical term that has found its way into the C-suite vocabulary is DevOps.
Within information technology, DevOps has its roots as far back as 2009 when the first devopsdays conference was held in Ghent, Belgium. By 2014, the State of DevOps report noted 30 times increased in the deployment of DevOps with 50% fewer failures.
FutureCIO spoke to Mike Kaczmarksi, IBM Fellow and CTO for IBM Management and Platform, to better understand how DevOps and microservices are evolving as organisations push further into their digital transformation.
A variant of the software-oriented architecture (SOA), microservices are a software development technique that breaks down an application into a series of small, more specialised software components designed to perform a specific task. Microservices communicate with each other using common interfaces like APIs and HTTP.
By breaking software into smaller chunks, it facilitates making micro changes to a code without affecting an application in its entirety. It also reduces the time it takes to introduce changes to an application.
DevOps defined
IBM’s Kaczmarksi says DevOps is about using enough automation in the development-build process so that the developers can release an application into production at any time. “It is a continuous building, continuous integration, continuous deployment and delivery, both in the test and staging with the idea that the software can be released into production,” he added.
Avoiding DevOps sprawl
The popularity of virtual machines (VMs) has resulted in a proliferation of VMs. Similarly, the accelerated use of containers to speed up development processes has inadvertently resulted in the unmonitored proliferation of microservices.
Kaczmarksi says microservices are services that you want to build and deploy. “At IBM cloud, we figure out what the service architecture is and we assign micro services for each of those services that are deployed in the architecture. We then build teams around the specific architectures. The team is tightly organized around the services provided it understands what each service needs to do. This way we avoid the sprawl,” he explained.
Role of AI, automation and ML in DevOps
Faster time-to-market is one of the key incentives driving the adoption of DevOps. One of the key ways to do this is through automation.
IBM’s Kaczmarksi acknowledged that automation is at the heart of DevOps. “AI comes in when you gather the information from a DevOps tool chain. You can use automation and analytics tools to identify errors, whether the function you are developing is good, how prepared are you to go into production, etc.,” he concluded.
Persistent misinterpretations
Using DevOps doesn’t mean that developers have a free reign on what they are writing or developing. It is not a return to the Wild West of old where coders are given freedom to develop their ideas – their way without care for the others.
“DevOps allows developers to see what their effort [applications they are working on] looks like in production,” noted Kaczmarksi.
The one question to ask
Managers and developers need to understand that for DevOps to work requires more than just adding tools that do version tracking, automate process, among others. Other considerations include making sure it is secure, does it support continuous deployment, and can the processes scale?
Asking the right questions at the onset may minimise surprises in the future.
DevOps best practices
Any DevOps best practice calls for an understanding of the build and deployment processes. It also involves plugging in the right capabilities into that process for the assurance. Security practices during DevOps can be used to understand whether the code exposes any vulnerabilities in the communication, noted IBM’s Kaczmarksi.
“One of the highest priorities I put on the DevOps process is to make sure you cook the security into the development into the build into the deployment, and continually evaluate your security posture relative the applications that are being deployed,” he concluded.
What to expect in 2020
IBM Fellow and CTO for IBM Management and Platform, Mike Kaczmarksi predicts that DevOps toolchains will be right into application development tools so that out-of-the-box, application developers can start writing their code, save their code, and automatically deploy into running environments. DevOps can enable developers to see what their changes right away.
DevOps will fail not because of tech
Brenda Christopher, Application Performance Management lead for IBM, provides one other cautionary tale for companies looking to implement DevOps.
“DevOps involves increasing collaboration and implementing a tighter feedback loop between the development and operations teams. This enhanced connectivity between the teams translates to an increased speed of delivery, along with increased reliability and stability in production,” she commented.
Gartner predicts that through 2022, 75% of DevOps initiatives will fail to meet expectations due to other than technology.
The increased collaboration referred to by Christopher may well be what George Spafford, senior director analyst at Gartner when he wrote: “Organizational learning and change are key to allowing DevOps to flourish. In other words, people-related factors tend to be the greatest challenges — not technology.”
