Reference Architecture: Building a Cost Optimized, High-Throughput HL7, EDI, and FHIR Processing Pipeline on OCI

Details

The document details an event-driven reference architecture deployed on OCI to manage unpredictable healthcare transaction volumes (e.g., claims intake, eligibility verification). The core principle is decoupling the transaction intake layer from downstream processing engines to ensure resilience during traffic spikes while maintaining cost efficiency. The pipeline handles multiple inputs: FHIR REST messages via API Gateway; real-time HL7 v2 messages received by Oracle Integration for Healthcare (using MLLP) which then transforms and stores the data in Object Storage; and batch files (X12 EDI, CSV) processed through OIC SFTP adapters. The processing flow is orchestrated by OCI Events triggered upon file upload to Object Storage, activating serverless OCI Functions. These functions publish normalized messages to OCI Streaming. Downstream stateless workers hosted on OKE consume these streams, allowing the system to scale based on stream lag and backlog. This buffered approach protects downstream applications from being overwhelmed during peak loads. By leveraging various OCI capabilities (API Gateway, Object Storage, Events, Functions, Streaming, OKE), the architecture achieves high throughput, scalability, durability, and cost optimization by minimizing unnecessary data movement and compute resource waste.