Akbar Travels
Global Booking Engine Architecture
Engineered a high-throughput travel booking engine processing 50,000+ daily searches across 900+ airline APIs with sub-second response times, intelligent fare caching, and multi-currency payment orchestration.

The Structural Challenge
Akbar Travels, one of India's largest travel agencies with 200+ branches nationwide, needed to transition from a primarily offline booking operation to a digital-first platform capable of competing with OTA giants like MakeMyTrip and Cleartrip. Their existing booking system was a vendor-locked desktop application that could not support web or mobile channels.
The technical challenge centered on aggregating real-time fare data from 900+ airline, hotel, and transportation APIs — each with different data formats, rate limits, and response characteristics. Some GDS (Global Distribution System) APIs had response times exceeding 8 seconds, making sequential queries to multiple suppliers impractical for a consumer-facing application targeting sub-second search results.
Currency conversion across 40+ currencies, complex fare rules with dynamic pricing, and the need to hold inventory across multiple suppliers simultaneously during the booking flow added layers of transactional complexity that demanded careful system design.
The Systems Architecture & Solution
We built a parallel search aggregation engine that dispatches queries to multiple GDS and direct-connect APIs simultaneously, with configurable timeout thresholds per supplier. A smart caching layer using Redis stores frequently searched route-date combinations with TTL values calibrated to fare volatility — popular routes refresh every 30 seconds while less-traveled routes cache for up to 15 minutes. This caching strategy serves 72% of searches from cache, dramatically reducing upstream API load.
The booking engine implements a distributed reservation system using the saga pattern to coordinate inventory holds across multiple suppliers. If any supplier fails during the multi-step booking process, compensating transactions automatically release held inventory on other suppliers, preventing orphaned bookings and ensuring accurate availability data.
Payment orchestration handles multi-currency transactions with intelligent gateway routing — selecting the optimal payment processor based on currency pair, transaction amount, and historical success rates. The system supports partial payments, EMI options, and corporate credit terms while maintaining PCI DSS Level 1 compliance.
Architecture Decisions
Parallel API aggregation engine with per-supplier timeout and fallback strategies
Tiered fare caching with volatility-based TTL reducing upstream API calls by 72%
Saga-based distributed booking with automatic compensating transactions
Multi-currency payment orchestration with intelligent gateway routing
The Measurable Enterprise Impact
The platform launched with immediate impact on Akbar Travels' competitive positioning in the Indian OTA market. Search response times averaged 780ms — comparable to market leaders — despite aggregating results from significantly more suppliers, giving users broader fare options. The booking success rate reached 99.4%, with failed bookings automatically retried through alternate suppliers.
Revenue grew 42% in the first year post-launch, driven by the new web and mobile channels reaching customers who previously booked through competitor platforms. The intelligent pricing engine, which factors in historical fare patterns and demand signals, improved average booking margin by 12% through optimized supplier selection.
800ms
Search Speed
Average response time aggregating 900+ supplier APIs in parallel
72%
Cache Hit Rate
Serving majority of searches from intelligent fare cache
+42%
Revenue Growth
First-year growth driven by web and mobile channel launch
+12%
Booking Margin
Improved through intelligent supplier selection and pricing optimization
“The platform gave us the technology to compete with the biggest OTAs while maintaining the supplier relationships that are our competitive advantage.”