AWS re:Invent 2025 - FIS: High-performance instant payment processing at massive scale (IND3318)

Transforming Payments at Scale: FIS and AWS Collaboration

Overview

• FIS, a leading global financial technology provider, collaborated with AWS to rethink payment solutions from the ground up and build a high-performance, massively scalable payment processing platform called the "Money Movement Hub".
• The session covers the key requirements, technical challenges, architectural decisions, and the innovative solutions implemented to address scalability, performance, resiliency, and observability.

The Changing Payments Landscape

• Payments have become an integral part of daily life, with increasing customer expectations for faster, more seamless, and always-on experiences.
• Traditional payment rails and batch-oriented banking systems are struggling to keep up with the demand for instant, programmable, and intelligent payment processing.
• FIS recognized the need to rethink payments from the ground up, moving away from siloed payment systems towards a unified, orchestrated, and intelligent payment processing platform.

Key Technical Requirements

1. Performance: Ability to execute and settle payments in under 5 seconds, meeting the requirements of instant and real-time payment schemes.
2. Scalability: Designed as a multi-tenant solution, capable of processing over 1,000 payments per second to handle fluctuating payment volumes.
3. Availability: Targeted 99.995% availability, allowing for less than 30 minutes of downtime per year.

Technical Challenges

• Seamless integration with existing on-premises banking systems and legacy technologies.
• Ability to easily extend the system to support new payment rails, such as digital currencies and stable coins, without a complete overhaul.
• Ensuring end-to-end observability across the complex payment processing pipeline.

Architectural Approach

• Adopted a multi-account strategy on AWS to isolate different subsystems and enable independent teams to work with different operating models.
• Implemented an event-driven, microservices-based architecture orchestrated by Kubernetes (EKS) for high scalability and resiliency.
• Leveraged AWS services for high-performance compute (EKS), database (Amazon Aurora), caching (Amazon ElastiCache), and data processing (Amazon MSK, AWS Glue).
• Utilized open-source technologies like Cadence, Conductor, and Kubernetes autoscalers to enhance scalability, performance, and workflow management.
• Designed for end-to-end observability using AWS CloudWatch, including the innovative use of CloudWatch Application Signals for service-level objective monitoring.

Key Technical Innovations

1. Scalability:

   • Used Kubernetes (EKS) with Cadence and Carpenter for dynamic pod and node scaling to handle fluctuating payment volumes.
   • Leveraged warm pools and custom VPC CNI plugins to ensure immediate pod schedulability and IP address management.

2. Performance:

   • Optimized container startup times using Bottlerocket and multi-stage Docker builds.
   • Utilized custom Spring annotations to route read queries to Aurora read replicas and asynchronous writes to the cache.
   • Evaluated Spring Ahead of Time and GraalVM for further performance improvements.

3. Resiliency:

   • Leveraged native cross-region replication capabilities of AWS services like Aurora, ElastiCache, and S3.
   • Used MSK Replicator to replicate Kafka topics and consumer offsets for seamless failover.
   • Implemented fault injection testing using the AWS Fault Injection Simulator.

4. Observability:

   • Adopted OpenTelemetry with the AWS distro to capture logs, metrics, and traces, and send them to a centralized observability account.
   • Utilized CloudWatch Application Signals to monitor service-level objectives (SLOs) and provide end-to-end visibility.

Business Impact and Outcomes

• Reduced time-to-market from concept to production from 9 months, a significant improvement over FIS's traditional payment system deployments.
• Enabled rapid deployment of new payment rails and capabilities, with 11 major initiatives planned for the next year.
• Empowered financial institutions with self-service capabilities and automated decision-making through the use of AI and machine learning.
• Facilitated the decommissioning of legacy payment systems and the migration of customers to the new Money Movement Hub platform.

Key Lessons Learned

1. Event-Driven Architecture: Start with an event-driven, microservices-based design to achieve massive horizontal scalability and resiliency.
2. Performance Optimization: Identify and optimize the critical real-time processing pipeline, leveraging high-performance compute, database, and caching services.
3. Resilient by Design: Leverage native multi-region and multi-AZ capabilities of AWS services to build a highly resilient solution.
4. Unified Observability: Implement end-to-end observability, including both technical and business-level metrics, to gain deep visibility into the system's health and performance.