Examkin

Question

Difficulty: MediumDesigning Migration Plans and Data Transfer Strategies

A global online travel booking platform is migrating its infrastructure to Google Cloud within a strict 3-week cutover window. The application stack includes a 15 TB15\text{ TB} operational PostgreSQL relational database that requires less than 2 hours of downtime during cutover, and a 350 TB350\text{ TB} historical search log archive. The enterprise maintains an existing 1 Gbps1\text{ Gbps} Dedicated Interconnect connection to Google Cloud, but network policies restrict migration bandwidth on this link to off-peak hours (maximum 12 hours per day). Which migration and data transfer strategy should the cloud architect recommend to meet all business and technical constraints?

  1. Order a Google Cloud Transfer Appliance to transfer the 350 TB350\text{ TB} historical search log archive offline, and use Database Migration Service (DMS) with continuous Change Data Capture (CDC) over the Dedicated Interconnect for the PostgreSQL database.Answer
  2. B
    Use Storage Transfer Service over the Dedicated Interconnect during off-peak hours to transfer both the 350 TB350\text{ TB} historical log archive and periodic PostgreSQL database backups.
  3. C
    Migrate the 15 TB15\text{ TB} PostgreSQL database to Cloud Spanner using Database Migration Service to eliminate cutover downtime, and transfer the search log archive via Storage Transfer Service over a temporary HA VPN connection.
  4. D
    Provision additional HA VPN tunnels to stream the 350 TB350\text{ TB} log archive using Storage Transfer Service, and perform a cold database migration using static database exports during the final maintenance window.

Answer

Order a Google Cloud Transfer Appliance to transfer the historical search log archive offline, and use Database Migration Service (DMS) with continuous Change Data Capture (CDC) over the Dedicated Interconnect for the PostgreSQL database.
Transferring 350 TB350\text{ TB} of data over a 1 Gbps1\text{ Gbps} line capped at 12 hours per day provides only ∼5.4 TB\sim 5.4\text{ TB} of daily ingestion, requiring over two months. Utilizing Google Cloud Transfer Appliance enables shipping the archive offline to meet the 3-week timeline constraint. Meanwhile, Database Migration Service (DMS) continuously replicates PostgreSQL transactions using Change Data Capture (CDC) over the Dedicated Interconnect, allowing the database cutover to complete well within the 2-hour target window.

Step-by-Step Solution

1
Calculate available online bandwidth capacity for the bulk log archive transfer.
At 1 Gbps1\text{ Gbps} (125 MB/s125\text{ MB/s}), 12 hours of transfer per day yields 125 MB/s×43,200 seconds≈5.4 TB/day125\text{ MB/s} \times 43,200\text{ seconds} \approx 5.4\text{ TB/day}. Transferring 350 TB350\text{ TB} requires approximately 65 days.
Because 65 days exceeds the 3-week (21 days21\text{ days}) timeline, online transfer via Dedicated Interconnect or HA VPN is unfeasible. An appliance-based offline transfer (Transfer Appliance) is required.
2
Evaluate relational database migration and cutover requirements.
A static dump/restore of a 15 TB15\text{ TB} database over the network or via appliance would require many hours to export, transport, and re-index, violating the 2-hour maximum downtime window.
Database Migration Service (DMS) utilizes continuous Change Data Capture (CDC) to keep the target Cloud SQL/PostgreSQL instance in sync with on-premises, enabling a rapid cutover within minutes.

Key Concept

Designing hybrid migration plans that combine offline physical appliances (Transfer Appliance) for bandwidth-constrained large datasets with continuous online replication tools (DMS with CDC) for low-downtime database cutovers.
Rate this question

Topics

Designing Migration Plans and Data Transfer Strategies