Modernizing Legacy Integration for a Leading Petroleum Enterprise
Replacing BizTalk middleware with a cloud-native Azure architecture capable of handling 2M+ daily API hits across 100+ customers and multiple file formats.
The Challenge
Legacy Middleware Couldn't Keep Up
A leading petroleum enterprise operated a mission-critical integration platform built on Microsoft BizTalk Server. The platform served as the central nervous system for data exchange with over 100 customers, processing more than 2 million API transactions daily.
The legacy BizTalk infrastructure was becoming a bottleneck. Scaling was expensive and time-consuming. Onboarding new customers or file formats required weeks of development. Monitoring was limited, and the operations team lacked real-time visibility into transaction health and failure patterns.
The client needed a modern, cloud-native integration platform that could handle the same volume with better reliability, faster onboarding, multi-format support, and real-time operational dashboards — all without disrupting existing customer integrations.
Our Solution
Disconnected Cloud-Native Architecture
We designed an asynchronous, event-driven architecture on Azure PaaS that decoupled ingestion from processing, enabling independent scaling and fault tolerance at every stage.
Receive & Validate
Incoming API requests from 100+ customers are received through Azure API Management (APIM). Each request is validated against schema rules, authentication, and rate limiting. Supports JSON, XML, CSV, and proprietary petroleum file formats.
Acknowledge & Queue
Upon successful validation, an immediate acknowledgment (ACK) is sent back to the client confirming receipt. The validated payload is then placed onto Azure Service Bus queues for reliable, asynchronous processing.
Transform & Process
Azure Function Apps pick up messages from Service Bus, transform data across formats (JSON, XML, CSV, proprietary), apply business rules, and prepare payloads for persistence. Azure Logic Apps orchestrate complex multi-step workflows.
Save & Balance
Transformed data is routed through Azure Load Balancer to the Save API layer. Based on current load and traffic patterns, requests are distributed across Kubernetes pods to ensure consistent performance at 2M+ daily hits.
Final Acknowledgment
Once data is successfully persisted, a final acknowledgment is sent back to the client API confirming successful processing. Failed transactions are routed to a retry queue with dead-letter handling.
Monitor & Report
Cosmos DB stores transaction metadata for real-time dashboards showing daily validation results, success/failure rates, and processing metrics with graph-based visualization for operations teams.
Technology Stack
Built on Azure PaaS
Outcomes
Key Results Delivered
Facing a Similar Challenge?
Whether it's legacy modernization, cloud migration, or building high-throughput integration platforms — let's talk about how we can help.
Book a Free Consultation