Fully-managed relational database service supporting MySQL, PostgreSQL, and MS SQL Server. Designed for structured, relational data.
Cloud SQL is a fully-managed relational database service. It is designed for relational, record-based data (tables, rows, and columns) and super structured data. It scales vertically and allows field updates.
SQL databases generally scale vertically for structured data. Note that no autoscaling for compute nodes is supported natively in Cloud SQL, though auto-scaling for storage exists.
As a fully-managed service, Google owns:
Provide details on your existing data source
Create a read replica in Cloud SQL
Sync the read replica with the source database
Promote the read replica to the primary instance
Cloud SQL is optimized for transactions (high throughput WRITES to RECORDS). It is record-based, meaning an entire record must be opened on disk even if you SELECT a single column. BigQuery, in contrast, is column-based and optimized for reporting workloads (mostly reads).
gsutil if copying files over from on-premise.The Cloud SQL Auth proxy offers enhanced security, especially for apps with dynamic IP addresses.
Q1. You have an application running in GKE with dynamic IP addresses that needs to connect to Cloud SQL securely. What is the best way to handle this?
Q2. Your analytics team needs to run heavy reporting queries, but doing so on Cloud SQL degrades performance for the web application. How can you solve this?
Q3. True or False: Cloud SQL automatically scales its compute nodes when CPU utilization hits 90%.
Q4. What is the fundamental difference between how Cloud SQL and BigQuery store data?