Planning and validating a migration of data and workloads into Google Cloud
Migrating data to Google Cloud requires choosing the right tool based on data source, volume, and latency requirements, then validating the migration for completeness and correctness. Google Cloud offers a spectrum of tools from offline hardware transfer to online continuous replication, and network planning (bandwidth, Dedicated/Partner Interconnect, VPNs) is critical for large-scale online migrations.
1 · Learn the must-know
- BigQuery Data Transfer Service (BQ DTS) automates recurring batch loads into BigQuery from SaaS apps (Google Ads, YouTube), other GCP services (Cloud Storage), and other data warehouses (Teradata, Redshift, Amazon S3), but it is scheduled/batch, not real-time.
- Database Migration Service (DMS) provides mostly-automated, minimal-downtime migration for MySQL, PostgreSQL, and SQL Server into Cloud SQL or AlloyDB, supporting both one-time and continuous replication migrations.
- Transfer Appliance is a physical, high-capacity storage device (rackable, up to hundreds of TB) shipped to your site for offline data capture and shipped back to Google for upload, ideal when network bandwidth is insufficient or transfer time via network would exceed practical limits (the classic 'sneakernet' tradeoff: compare estimated network transfer time vs. shipping time).
- Storage Transfer Service moves data online between object storage systems (on-prem POSIX filesystems, other clouds like S3/Azure Blob, or between GCS buckets) and supports scheduled, incremental syncs with filtering, distinct from BQ DTS which targets BigQuery specifically.
- Datastream is a serverless CDC (change data capture) and replication service that streams changes in near real-time from Oracle, MySQL, and PostgreSQL sources into destinations like BigQuery or Cloud Storage, commonly paired with Dataflow templates for downstream processing.
- For large or ongoing online migrations, plan network capacity explicitly: use Cloud Interconnect (Dedicated or Partner) or Cloud VPN for private, higher-bandwidth connectivity rather than relying on public internet egress, and always validate migrated data via row counts, checksums, or sampling before cutover.
2 · Check your understanding
A retail company runs a MySQL database on-premises and wants to replicate change data continuously into BigQuery for near real-time analytics, with minimal impact on the source database and minimal downtime during cutover. Which Google Cloud service should they use to capture and stream the change data?
What you have tried across GCP PDE's objectives, not a readiness score.
Designing data processing systems~22% of the exam0 of 17 tried
Ingesting and processing the data~25% of the exam0 of 11 tried
Storing the data~20% of the exam0 of 13 tried
Preparing and using data for analysis~15% of the exam0 of 11 tried
Maintaining and automating data workloads~18% of the exam0 of 15 tried
3 · Keep going
Ready for more? Take a weighted mock or try free practice questions.