Transforming payment processing for banking channels

Multiple banking channels - mobile, internet and branch - initiate various payment requests (20+ types such as ACH, mandate and book transfer) that need to be processed. Depending on the payment type, the channel was required to invoke one or more services in a specific order as per the associated business rules.

Changes to these payment workflows stemming from introduction of new payment types or revisions of business rules introduced complex and repeated changes to the bank’s systems, hindering scalability.

• To support lifecycle management of various payment transactions, including defining different payment workflows, we designed an event-driven architecture comprised of 40+ microservices (e.g., limit, eligibility, and fraud checks, etc.) supported by a Kafka message queuing system
• The architecture involved building an orchestration engine (landing service), acting as a front controller for all payment workflow requests from the various banking channels, such as mobile, internet and branch
• The landing service in turn invokes the corresponding service (limit, eligibility, etc.) based on the payment type and business rules associated with it.
• Data flow between these microservices (resulting from further invocations) and other downstream systems is facilitated asynchronously with the help of a distributed messaging system (Kafka)
• Using Jenkins, we built a CI/CD pipeline to streamline the workflow by automatically building, testing and deploying code changes as they are committed

• Significantly eased the management of payment workflows, including those related to the addition of new payment types (resulting from an acquisition)
• Enabled systems to scale without introducing complex changes at the channels
• Improved reporting, resulting from faster access to data through dedicated microservices