Title

AWS re:Invent 2023 - Bringing workloads together with event-driven architecture (API206)

Summary

Dheeraj Mahapatra, a principal serverless persist SA at AWS, and Nick Stompos from JPMorgan Chase discussed event-driven architecture (EDA) and its applications.
Nick shared JPMorgan's journey from mainframe-based systems to modern EDA for merchant services, highlighting the need for a global, secure, and configurable solution.
Dheeraj demonstrated an insurance claims processing application using AWS services to showcase EDA's benefits, including decoupling, asynchronous communication, and scalability.
The session covered various AWS services and patterns for EDA, including EventBridge, Lambda, Step Functions, Fargate, EKS, and KEDA.
Dheeraj emphasized the extensibility, polyglot support, and reduced impact area of EDA, and how it enables seamless integration of different workloads and compute options.
The talk concluded with a call to action for further learning through AWS Skill Builder and visiting the re:Invent Expo for more examples and resources.

EDA is crucial for modernizing legacy systems, as demonstrated by JPMorgan's transition from mainframe-based batch processing to a more flexible, scalable, and resilient cloud-based architecture.
AWS offers a rich ecosystem of services that facilitate EDA, including serverless options that reduce operational overhead and allow for fine-grained scaling.
The use of EDA can significantly reduce the impact of changes and faults by isolating services and domains, allowing for independent scaling and evolution of each component.
Hexagonal architecture is a beneficial design pattern for implementing EDA, as it abstracts core business logic from the communication mechanisms, making it easier to adapt to event-driven models.
The session highlighted the importance of understanding the nature of EDA, including asynchronous events, event brokers, and event stores, to effectively leverage AWS services for building event-driven systems.
The use of AWS CDK and infrastructure as code tools like KEDA can streamline the integration of containerized workloads with EDA, providing auto-scaling capabilities based on event triggers.
The talk underscored the potential for EDA to integrate with various compute options, including Lambda, Fargate, and EKS, showcasing the flexibility of AWS in supporting diverse application needs.