Nationwide Scales Amazon EKS to Simplify and Accelerate IT Development

Headquartered in Swindon, England, Nationwide provides a range of banking and financial services to its members, including loans, credit cards, insurance, and investments. In 2017, the institution’s leadership team embarked on a digital transformation journey to strengthen its internal engineering capabilities, adopt new ways of working, simplify its technology footprint, and reduce operational complexity.

As part of this initiative, different teams began to experiment with cloud technologies. As Nationwide observed these efforts, it recognized the advantages of the cloud: teams can quickly deploy new workloads, accelerating development cycles.

“This spurt of innovation and experimentation increased efficiency in pockets, but it lacked a strong enterprise focus,” says Dave Burke, CTO at Nationwide. “Moreover, this also introduced risks and challenges from a controls and governance perspective.”

To standardize these processes, Nationwide began to build ePaaS. It adopted Kubernetes, an open-source container orchestration system, to deploy containers for business workloads. Initially, teams built bespoke Kubernetes-based solutions, but these processes were difficult to scale and manage.

To simplify the ePaaS environment and reduce operational overhead, Nationwide recommissioned the build of secure, centralized, and cloud-native ePaaS models on AWS. It adopted Amazon Elastic Kubernetes Service (Amazon EKS), the most trusted way to start, run, and scale Kubernetes, to facilitate the federated deployment of its clusters.

With ePaaS, Nationwide standardized its cloud approach while reducing operational overhead. Under the Shared Responsibility Model, AWS is responsible for managing the controls associated with the infrastructure deployed in the AWS environment. Nationwide can configure Amazon EKS as needed to meet its security and governance requirements. It no longer needs to manage container orchestration, update deployment, or workload optimization.

Engineering teams rely on Amazon EKS to handle the heavy lifting of security and operating Kubernetes at scale. The centralized ePaaS integrates with Nationwide’s AWS landing zone, as well as other cloud-native services for monitoring, logging, secrets management, and other capabilities. As a result, developers no longer need to build network and security integrations themselves. And by adopting Amazon Elastic Compute Cloud (Amazon EC2) Spot Instances, which run fault-tolerant workloads for up to 90 percent off, Nationwide immediately reduced its total cost of ownership.

“A standardized ePaaS reduced our duplication of effort. It is deployed in a decentralized manner, yet managed and maintained centrally,” says Tim Wright, head of cloud operations at Nationwide. “The operational overhead of a self-managed cluster has virtually gone away, which lets us report on compliance against controls across the organization.”

To develop ePaaS, Nationwide implemented best practices, standards, and blueprints created by an internal group. Key principles such as “everything as code” drove further standardization and consistency. Nationwide also adopted chaos engineering, a practice that can help financial services institutions bridge the operational resilience gap. Chaos engineering involves conducting experiments to test the failure resilience of distributed software systems. By doing so, it helps identify system weaknesses. This process is crucial for demonstrating to regulators the capability of production-scale distributed software systems to withstand turbulent phases. With chaos engineering, Nationwide successfully tested and validated resilience for ePaaS-based workloads.

To encourage ePaaS adoption, Nationwide established a Cloud Engineering Team and used the Shared Responsibility Model to define the roles and responsibilities of different teams. To ease the onboarding for important initial consumers, it embedded platform engineers in product teams, scaling ePaaS skills across its organization. Nationwide hosted several AWS GameDays, which are collaborative learning exercises that test skills in implementing AWS solutions, to further improve operational maturity.

Nationwide also implemented several features to facilitate self-service for ePaaS users, including automated patch/feature releases and a postbuild stage where developers can incorporate their own continuous integration/continuous deployment toolchains.

“By moving to a federated self-service model, we helped grow the adoption of cloud technologies within Nationwide,” says Dave Burke. “Within 18 months of being generally available, ePaaS is used by over 15 teams, and there are over 150 Amazon EKS clusters in operation. On AWS, teams can focus less on provisioning and maintaining their infrastructure and more on delivering new, innovative products to members.”

Now, ePaaS hosts critical applications for Nationwide, including mortgage applications, affordability calculators, and credit card originations. The company is actively incorporating inputs from ePaaS consumers to drive the solution’s road map. As it moves forward, this feedback will help ePaaS meet even more of the institution’s business objectives and the needs of its members.

“Amazon EKS has facilitated numerous migrations from legacy cloud systems to ePaaS, providing increased security, resiliency, and scalability to our internal consumers,” says Tim Wright. “In the future, we will expand its hosting potential by supporting more critical business applications.”