Enterprise · 2023-2026 — Booking.com Rides
BDX Migration & Modernisation.
Multi-year data infrastructure modernisation for the Rides vertical at Booking.com. Hadoop-to-Snowflake migration spanning 200+ data sources, 80+ workflows, and 12 legacy systems — completed on target in April 2026.
209
Workflows
342
Assets
12
Legacy Systems
April 2026
Completed
Overview
Led the Rides data infrastructure modernisation programme at Booking.com, moving the entire vertical off legacy Hadoop onto Snowflake (BDX). The programme touched 209 workflows, 342 assets, 12 legacy systems, and 11 Tableau reports across Data Engineering, DSAA, FP&A, TRIM, Legal, and Compliance. The end-of-April 2026 target was hit: 100% of in-scope workflows migrated, all assets validated, Tableau reports live over BDX, and legacy DWH decommissioning underway.
Scope
209
Total Workflows
100% migrated on target
342
Total Assets
216 deprecated / decommissioned
10
Glue Jobs
Kinesis → BDX
11
Tableau Reports
Consolidated and re-validated
What Made This More Than a Migration
- —New data contracts established between Data Engineering, DSAA, and FP&A — no more implicit dependencies.
- —Business logic rebuilt from scratch (not lift-and-shift) — pipelines reflect current domain understanding, not legacy assumptions.
- —Scalable pipelines on modern cloud-native infrastructure, replacing brittle Hadoop workflows.
- —Modernised reporting — 11 Tableau reports consolidated and re-validated over BDX.
- —Simultaneous C360 integration replacing behavioural event streaming from legacy systems.
- —10 AWS Glue Jobs migrated from Kinesis to BDX as part of the same programme increment.
Programme Timeline
Legacy DWH Shutdown — 5 Steps
Notify & migrate active users
Communicate cut-over timeline to all active users on the legacy DWH. Coordinate migration paths to BDX for each team.
Validate reports, dashboards & assets
Run parallel validation across all Tableau reports and downstream assets — confirm output parity between legacy DWH and BDX.
Final checks & parallel run
Execute a full parallel run with zero-data-loss guarantee. Sign-off from Data Engineering, DSAA, and FP&A before cutover.
Offboard users & cut DB ties
Remove all active connections to the legacy DWH. Revoke credentials, update connection strings across all consumers.
Complete decommissioning
Shut down legacy infrastructure. Archive historical snapshots per Legal and Compliance requirements. Close out the programme.
Key Decisions & Risks
- —Rebuild over lift-and-shift: every pipeline was rewritten with current business logic to eliminate years of accumulated debt.
- —Parallel run period mandated for all high-criticality workflows before cutover — data loss risk effectively zeroed.
- —216 assets deprecated during migration, not carried forward — explicit decision to not migrate dead weight.
- —C360 integration run concurrently with BDX migration, sharing the same cutover window to minimise disruption.
- —Cross-functional sign-off from Legal and Compliance required before any legacy system decommissioning — no shortcuts.
- —Tableau reports validated by FP&A over BDX before legacy access was revoked — no business interruption.