Javed Roshan, Head of Data Services at Capital One, talked about how his company is using container technology.
In his keynote presentation at the 2016 Chief Information Leadership Forum held in San Francisco on February 18, Roshan started out by stating that about 90% of Capital One’s Cloud, fast-data platform is built on containers. His company is looking to migrate its on-premises infrastructure onto the Cloud, and containers have been essential to do this. Containers manage the infrastructure. The company has adopted DevOps as a model of operation, and containers can help with defining the operating model of an organization, said Roshan.
Capital One is adopting microservices architecture as a technology, and containers are a natural fit with this architecture, said Roshan. The company was formed on the basis of data analytics, even before it was called that, back in the 1990s. Containers help to create new algorithms to define the company’s lending practices.
Capital One uses containers in development. “We hear developers or engineers say a particular functionality doesn’t work in this case or in a particular environment. There are two reasons for this—the environments aren’t consistent or there’s an issue with data,” explained Roshan. The data issues can be solved relatively easily, but making environments consistent is more of a challenge. Container technology has helped Capital One create consistent environments.
Containers are also used in testing technologies.
“Nowadays, containers available with any product or technology have all the dependencies that are needed built into it. You simply go get it on a developer’s laptop and start working with it,” said Roshan.
“Nowadays, containers available with any product or technology have all the dependencies that are needed built into it. You simply go get it on a developer’s laptop and start working with it.”
The same is true for services that Capital One is developing. “We package all our services along with the dependencies so we can easily deploy the service,” explained Roshan. “The service can’t function in isolation.”
Another advantage of containers is that they’re making developers more agile. Containers allow them to be more innovative and not reluctant to try any technology.
Continuous integration is another use for containers, and this is a great place to start for those who haven’t adopted container technology, observed Roshan. Capital One has an automated “build” process that’s containerized. “The build server is containerized and can be triggered on an event. The container has the intelligence to retrieve the code from the correct source code depository, and it builds the code. It also clears certain artifacts and pushes those into a registry. Capital One has created a centralized repository, a registry, from which all teams and organizations can go to manage their containers.”
Continuous delivery is another area that benefits from containers. Deployment is a mirror image of the build process, explained Roshan, and can be triggered on an event. The needed images are retrieved from a common registry and deployed. “Containers are going to minimize or eliminate the need for a virtual machine going forward. You can use this same process on premises or in the Cloud,” said Roshan.
“Containers are going to minimize or eliminate the need for a virtual machine going forward. You can use this same process on premises or in the Cloud.”
Roshan offered a few observations about containers going forward. As the secondary technology matures, he said, Windows is going to provide products and services with container-enabled technology. The next big phase in container technology is to extend it beyond the data platform, such as into business support systems.
ABOUT JAVED ROSHAN:
Javed Roshan, Director of Data Services, leads the modernization of Fast Data infrastructure at Capital One Financial Services. In his 20 years of experience in various technology leadership positions, he has developed innovative enterprise solutions, highly scalable cloud-based systems, containerized and distributed micro-services with a focus on real-time processing of data at scale.
He specializes in practicing ‘information-based strategy’ to solve complex business problems, product strategy and roadmap development, systems and application security, legacy systems transformation, and product integration.