Rebooking Engine
During a mass disruption (such as a severe weather storm closing a hub), finding and securing alternative flights for thousands of stranded passengers is a major bottleneck. The Nexa Rebooking Engine automates the search, generation, and locking of alternative flight options, allowing operators to present viable routes to passengers and enabling passenger self-service selection.
The engine handles three major challenges: connection-aware routing, seat-inventory holds, and the passenger decision loop.
Connection-Aware Routing
Simple point-to-point flight matching is insufficient for complex hub networks. The Rebooking Engine runs a connection-aware search pipeline that evaluates:
- Multi-Leg Itineraries: Generating routing options that involve connections or alternative layovers if direct flights are full.
- Flight Risk Signals: Interfacing with the platform's flight predictor to analyze the delay and cancellation risk of the suggested alternative flights before offering them.
- Entitlement Policies: Filtering flight options based on versioned airline policies (e.g., reserving business-class inventory for elite loyalty tiers).
Seat-Inventory Holds
To prevent multiple operators or passengers from booking the same seat simultaneously, the engine implements a soft-hold seat reservation pattern:
Key Resilience Properties:
- Self-Healing Expiry (TTL): When an option is generated, a transient lock is created in the coordination store with a time-to-live (TTL) of 20 minutes. If the passenger or operator abandons the flow, the lock expires automatically, releasing the seats back into the available pool.
- Optimistic Concurrency: The rebooking database enforces version checking on all decisions. If a concurrent event occurs (e.g., a seat is taken in the PSS before the hold completes), the write fails safely, preventing double-bookings.
The Passenger Decision Loop
Nexa gives passengers agency by letting them accept or decline rebooking offers directly from their mobile devices. This eliminates the need for passengers to stand in long airport queues.
- Offer Generation: When the operator (or an automated policy) triggers flight generation, the rebooking options are compiled and locked.
- Passenger Review: The passenger receives a link on their mobile WebApp. They see the route details, layovers, and estimated arrival times.
- Outcome:
- Accept: The system completes the booking in the PSS asynchronously and issues the new tickets.
- Decline / Timeout: The soft-holds are instantly cancelled, the seats are returned to the inventory pool, and the passenger is routed back to the operator dashboard for manual rework.
What's next
- Case Lifecycle — how rebooking decisions integrate with the case saga.
- Real-time Collaboration — how Nexa coordinates multiple operators.
- Data Model — the rebooking URN types.