---
title: "GDS vs Bedbank vs Direct Connect: Distribution Topology Guide"
description: "Compare GDS, bedbank/B2B marketplace and direct-connect models across inventory ownership, coverage, booking lifecycle and operations."
slug: "gds-bedbank-direct-connect"
translationKey: "guide-gds-bedbank-direct-connect"
locale: "en"
type: "guide"
category: "architecture"
tags: ["travel-distribution","architecture","canonical-model"]
publishedAt: "2026-09-26"
updatedAt: "2026-09-26"
reviewedAt: "2026-09-26"
technicalVerifiedAt: "2026-09-26"
codeExampleStatus: "illustrative"
sources:
  - title: "IATA — NDC / airline distribution"
    url: "https://www.iata.org/en/programs/airline-distribution/retailing/ndc/"
  - title: "HBX Group — Hotels API Suite"
    url: "https://developer.hotelbeds.com/documentation/hotels/"
---
GDS, bedbank and Direct Connect are not equivalent product types. A GDS is a multi-party distribution network; a bedbank distributes contracted or aggregated accommodation inventory to B2B buyers; Direct Connect is a topology between two parties.

## Inventory and commercial ownership

A bedbank may expose its own contract and rate semantics. A GDS transports supplier content through a distribution network. In Direct Connect, commercial and technical contracts are maintained directly between the parties.

## Search and booking

HBX Hotel Booking API demonstrates availability → optional CheckRates → booking. In a direct API integration this lifecycle is managed explicitly by your adapter. GDS booking, ticketing and servicing vary by provider capability and vertical.

## Decision matrix

Evaluate coverage, negotiated rates, content richness, latency, certification, credit/payment model, support and servicing needs together.

## Failure modes

Stale rates, mapping mismatches, supplier timeouts, recheck failures, unknown booking state and cancellation reconciliation can occur in all models, but ownership lands in different places.

## Observability

Track source/provider success rate, search and booking latency, offer freshness, mapping coverage, booking UNKNOWN rate, reconciliation drift and source-level cancellation/refund failures.

## Production checklist

- persist commercial ownership explicitly,
- preserve source-specific identities/references,
- separate search and booking contracts,
- use reprice/availability validation,
- define provider-specific timeout/idempotency,
- provide failure/reconciliation paths,
- operate source-level observability.

## Architecture rule

Keep the canonical Offer model independent of distribution topology and isolate supplier contracts inside adapters.