Major Airline Application Migration and Re-Architecture

Leverage highly available infrastructure to mitigate operational costs of maintenance and recurring outages that result in a loss of revenue loss, barring crew from being able to make connecting flights.

As part of a “Lift, Shift, and Tinker” exercise, migrate existing services to AWS Cloud in a dual region (active-active) deployment to minimize downtime due to outages while increasing functionality, resilience, security and observability.

The airline’s on-prem service was built on a monolithic architecture, lived without redundancy, and required a 4 hour deployment window when updates needed to be made. This prohibited 24/7 operations, and left zero room for issues that would cause an outage. 

As part of the modernization effort, the entire existing workload looked at through the lens of re-architecture. The team focused on breaking the monolith down into standalone micro-services to allow for pieces to be updated and improved.Utilizing multiple AWS regions with latency-based routing ensured that all micro-services have redundancy in the event of a total region outage. Finally, we updated deployments to follow Blue-Green deployment patterns to allow for releases and improvements around the clock without disrupting 24/7 operations.

This was accomplished in multiple iterations; the first pass was simply to lift and shift into the cloud, then to bring high availability, then blue-green deployments.

Micro-services were first built with API gateways to allow the service to exist and function, knowing that we would move to AppSync to allow for event driven designs to various front-ends, but to leave the APIs in place for external (in customer, but outside of the business group) consumers we knew we would need to share data with.