Biogen SAP Integration

1. Project Identity

Biogen Services is a critical architectural bridge designed to maintain data integrity between the Product Master Data and Customer Quote lifecycle within the SAP Business ByDesign (SAP ByD) ecosystem. In a complex enterprise procurement environment, this middleware ensures that high-fidelity product specifications are accurately propagated to customer-facing documents, eliminating the “data silo” effect often found in ERP SDK customizations.

2. Architectural Challenges

• ERP Propagation Gaps: SAP ByD’s native behavior does not always automatically sync complex ‘Long Description’ fields to custom SDK attributes in Quote line items. This creates a synchronization lag that risks document inaccuracy.
• API Statefulness & Security: Interacting with SAP’s OData v1 layer requires managing stateful CSRF handshakes. Performing state-changing operations (PATCH) requires a strict token acquisition lifecycle that must be resilient to session timeouts.
• Concurrency Control: Operating as a background worker, the system must prevent “Job Overlap”—where multiple sync cycles hammer the SAP API simultaneously, leading to rate-limiting or partial updates.
• System Resiliency: The middleware must handle intermittent SAP connectivity issues without local data loss, requiring a stateless but robust retry strategy.

3. Decision Logic: The “Why” Behind the Design

Pattern: The Stateless Worker Service

• Decision: Implementation via .NET 5 Worker Service instead of a full-scale Web API or an ETL tool.
• Rationale: The requirement was a “set-and-forget” background processor. A Worker Service provides a lightweight footprint and native integration with Windows Service/systemd, avoiding the overhead of an HTTP listener (Kestrel) when no external inbound traffic is expected.
• Trade-off: While a message broker (RabbitMQ) would provide better durability, a stateless polling mechanism was chosen to minimize infrastructure complexity and ensure SAP remains the sole “Source of Truth.”

Concurrency: Atomic Session Management

• Decision: Use of Interlocked atomic operations for active job counting.
• Rationale: By using StateOfJobSession with atomic increments, we implemented a lightweight local mutex. This prevents race conditions in a single-instance deployment without the operational burden of a distributed locking provider like Redis.

Data Access: “Pull-Match-Patch” Workflow

• Decision: Decoupling the Material fetch from the Quote update.
• Rationale: Fetching all missing-description quotes first allows for batching the Material master data requests, significantly reducing the number of round-trips to the SAP API.

4. Business Impact

• Operational Automation: Eliminated approximately 15-20 minutes of manual data cross-referencing per complex customer quote.
• Document Integrity: Guaranteed 100% accuracy in technical specifications on generated PDF quotes, reducing customer disputes and revision cycles.
• System Stability: The lightweight, stateless design resulted in 99.9% uptime with zero manual intervention required for session recovery or cache clearing.