A financial services institution is migrating its reporting warehouse and transactional wo… — Google Cloud Professional Cloud Architect

A financial services institution is migrating its reporting warehouse and transactional workloads from an on-premises data center to Google Cloud. The dataset consists of $1.2\text{ PB}$ of historical raw analytics logs stored on Network-Attached Storage (NAS) and an active $4\text{ TB}$ operational MySQL database. The institution has a dedicated WAN bandwidth of $200\text{ Mbps}$ available for migration activities, and business requirements mandate a maximum downtime window of 4 hours during the final cutover.

Which TWO architectural decisions should the Cloud Architect incorporate into the migration plan to satisfy these requirements?

Order GCP Transfer Appliance devices to perform an offline bulk data transfer of the $1.2\text{ PB}$ historical analytics logs into Cloud Storage.Answer
Use Database Migration Service (DMS) with continuous Change Data Capture (CDC) to replicate the $4\text{ TB}$ operational MySQL database prior to cutover.Answer
C
Stream the $1.2\text{ PB}$ historical analytics dataset directly into Cloud Storage over the $200\text{ Mbps}$ network link using Storage Transfer Service.
D
Migrate the operational MySQL database directly to Cloud Spanner to ensure single-region transactional consistency without requiring schema refactoring.

Answer

The correct architecture requires utilizing GCP Transfer Appliance for the offline bulk transfer of the

1.2\text{ PB}

historical files and Database Migration Service (DMS) with Change Data Capture (CDC) for continuous database synchronization prior to final cutover.

For the

1.2\text{ PB}

raw analytics logs, an online transfer over

200\text{ Mbps}

would take over 550 days. Transfer Appliance provides an offline physical hardware shipping model that ingests petabyte-scale data into Cloud Storage efficiently. For the active

4\text{ TB}

MySQL database, Database Migration Service (DMS) with CDC continuously streams change logs to Cloud SQL in real time, allowing a seamless cutover within the 4-hour window.

Step-by-Step Solution

Calculate network throughput feasibility for the bulk historical dataset transfer

200\text{ Mbps}

, transferring

1.2\text{ PB}

(

1.2 \times 10^{15}\text{ bytes}

) requires

\frac{1.2 \times 10^{15} \times 8}{200 \times 10^6 \times 86400} \approx 555\text{ days}

, confirming that an online transfer is impossible.

Petabyte-scale datasets on limited bandwidth connections mandate an offline physical appliance ingestion model using Transfer Appliance.

Evaluate operational database migration and cutover constraints

Continuous replication via Database Migration Service (DMS) with CDC performs the bulk load online and keeps the destination synchronized until cutover.

DMS minimizes cutover time down to a simple DNS/connection swap, easily satisfying the 4-hour downtime constraint.

Key Concept

Selecting appropriate GCP data migration tools based on network bandwidth calculations, dataset size, and database downtime windows.

Question

Answer

Step-by-Step Solution

Key Concept

Common Mistakes

Topics