Bookaweb
Multi-vertical travel marketplace operating across 210 countries.
A booking platform integrating heterogeneous inventory sources - partner-driven hotel and activity feeds alongside fully controlled direct property listings - unified under a shared search, transaction, and background processing infrastructure.
- 2.4M+ hotels
- 300K+ activities
- On-platform booking & payment
- Deterministic rule-based pricing engine (direct listings)
Project Summary
Bookaweb integrates multiple inventory domains - hotels, activities, and direct listings - into a single marketplace surface.
Each vertical operates as an independent bounded context with isolated data models, distinct booking semantics, dedicated ingestion and synchronization pipelines, and separate operational boundaries.
Partner-driven verticals ingest external pricing and availability under rate limits and partial upstream failure. Inventory is normalized and stored locally to ensure consistent query behavior and remove real-time dependency on partner APIs during search and checkout.
Direct listings run on a fully controlled transactional system with deterministic pricing evaluation, reservation lifecycle modeling, and conflict-safe state transitions.
Infrastructure (deployment, queues, search layer) is shared, while business logic remains isolated per vertical.
🧠 Engineering Highlights
- Vertical Domain Isolation: service-layer separation per vertical to prevent cross-domain coupling and reduce blast radius during ingestion spikes or partner outages.
- Deterministic Pricing Engine (Direct Listings): rule-compounding pricing engine with strict evaluation ordering and idempotent recomputation. Prices are precomputed to shift complexity off the read path and maintain low-latency filtering under high-cardinality queries. Fully test-driven.
- Multi-Source Ingestion & Synchronization: concurrent partner ingestion pipelines with rate limiting, staggered scheduling, retry handling, and integration-level failure isolation. Calendar synchronization reconciles external availability with internal reservations and owner constraints.
📊 Scale & Production Characteristics
- 2.4M+ hotel properties
- 300K+ activities
- 210-country inventory coverage
- Continuous multi-partner synchronization
- Precomputed pricing optimized for read-path latency
- Production transactional booking lifecycle
- Domain-segmented background processing
👤 Architectural Responsibilities
- Cross-vertical domain modeling and boundary enforcement
- Direct-listing pricing system design
- External API ingestion architecture
- Calendar synchronization and conflict detection modeling
- Reservation lifecycle state machine design
- Background workload segmentation and scaling
