Cloud SQL

Cloud SQL is a web service that allows you to create, configure, and use relational databases that live in Google’s cloud. It is a fully-managed service that maintains, manages, and administers your databases, allowing you to focus on your applications and services.

Google Cloud Storage Services Cheat Sheet

~ Last updated on : ~ jayendrapatil ~ 2 Comments

Google Cloud Storage Options

  • Relational (SQL) – Cloud SQL, AlloyDB & Cloud Spanner
  • Non-Relational (NoSQL) – Firestore & Bigtable
  • Structured & Semi-structured – Cloud SQL, AlloyDB, Cloud Spanner, Firestore & Bigtable
  • Unstructured – Cloud Storage
  • Block Storage – Persistent Disk & Hyperdisk
  • File Storage – Filestore
  • In-Memory – Memorystore (Redis, Valkey)
  • Transactional (OLTP) – Cloud SQL, AlloyDB & Cloud Spanner
  • Analytical (OLAP) – Bigtable & BigQuery
  • Fully Managed (Serverless) – Cloud Spanner, Firestore, BigQuery, AlloyDB
  • Requires Provisioning – Cloud SQL, Bigtable
  • Global – Cloud Spanner
  • Regional – Cloud SQL, AlloyDB, Bigtable, Firestore

Google Cloud - Storage Options Decision Tree

Google Cloud Storage – GCS

  • provides service for storing unstructured data i.e. objects
  • consists of bucket and objects where an object is an immutable piece of data consisting of a file of any format stored in containers called buckets.
  • support different location types
    • regional
      • A region is a specific geographic place, such as London.
      • helps optimize latency and network bandwidth for data consumers, such as analytics pipelines, that are grouped in the same region.
    • dual-region
      • is a specific pair of regions, such as Finland and the Netherlands.
      • provides higher availability that comes with being geo-redundant.
    • multi-region
      • is a large geographic area, such as the United States, that contains two or more geographic places.
      • allows serving content to data consumers that are outside of the Google network and distributed across large geographic areas
      • provides higher availability that comes with being geo-redundant.
    • Objects stored in a multi-region or dual-region are geo-redundant i.e. data is stored redundantly in at least two separate geographic places separated by at least 100 miles.
  • Storage class affects the object’s availability and pricing model
    • Standard Storage is best for data that is frequently accessed (hot data) and/or stored for only brief periods of time.
    • Nearline Storage is a low-cost, highly durable storage service for storing infrequently accessed data (warm data)
    • Coldline Storage provides a very-low-cost, highly durable storage service for storing infrequently accessed data (cold data)
    • Archive Storage is the lowest-cost, highly durable storage service for data archiving, online backup, and disaster recovery. (coldest data)
  • Autoclass automatically transitions objects to appropriate storage classes based on access patterns, removing the need to manually manage lifecycle rules for cost optimization. Supports transitioning between Standard, Nearline, Coldline, and Archive classes.
  • Soft Delete (launched 2024) provides default bucket-level protection against accidental or malicious deletion by preserving recently deleted objects for a configurable retention period (7-90 days). Enabled by default with 7-day retention.
  • Object Versioning prevents accidental overwrites and deletion. It retains a noncurrent object version when the live object version gets replaced, overwritten or deleted
  • Object Lifecycle Management sets Time To Live (TTL) on an object and helps configure transition or expiration of the objects based on specified rules for e.g. SetStorageClass to change the storage class, delete to expire noncurrent or archived objects
  • Resumable uploads are the recommended method for uploading large files, because they don’t need to be restarted from the beginning if there is a network failure while the upload is underway.
  • Parallel composite uploads divides a file into up to 32 chunks, which are uploaded in parallel to temporary objects, the final object is recreated using the temporary objects, and the temporary objects are deleted
  • Requester Pays on the bucket that requires requester to include a billing project in their requests, thus billing the requester’s project.
  • supports upload and storage of any MIME type of data up to 5 TB in size.
  • Retention policy on a bucket ensures that all current and future objects in the bucket cannot be deleted or replaced until they reach the defined age
  • Retention policy locks will lock a retention policy on a bucket and prevents the policy from ever being removed or the retention period from ever being reduced (although it can be increased). Locking a retention policy is irreversible
  • Bucket Lock feature provides immutable storage on Cloud Storage
  • Object holds, when set on individual objects, prevents the object from being deleted or replaced, however allows metadata to be edited.
  • Signed URLs provide time-limited read or write access to an object through a generated URL.
  • Signed policy documents helps specify what can be uploaded to a bucket.
  • Cloud Storage supports encryption at rest and in transit as well
  • Cloud Storage supports both
    • Server-side encryption with support for Google managed, Customer managed and Customer supplied encryption keys
    • Client-side encryption: encryption that occurs before data is sent to Cloud Storage, encrypted at client side.
  • Cloud Storage operations are
    • strongly consistent for read after writes or deletes and listing
    • eventually consistent for granting access to or revoking access
  • Cloud Storage allows setting CORS configuration at the bucket level only

Cloud Storage Rapid (New – 2025)

  • Rapid Bucket (formerly Rapid Storage) is a high-performance zonal object storage offering designed for AI/ML and HPC workloads
    • Provides sub-millisecond random read and write latency
    • Up to 15 TB/s of aggregate throughput and 20 million QPS
    • Collocates data with AI accelerators (TPUs/GPUs) in the same physical zone
    • 5x faster checkpoint restores and 3.2x faster checkpoint writes
  • Rapid Cache (formerly Anywhere Cache) accelerates reads on-demand for workloads in existing multi-region buckets, providing up to 20 Tbps throughput
  • Storage Intelligence provides zero-configuration dashboards, aggregated activity views, and enhanced batch operations for streamlined data management

Cloud SQL

  • provides relational MySQL, PostgreSQL and SQL Server databases as a service
  • managed, however, needs to select and provision machines
  • supports automatic replication, managed backups, vertical scaling for read and write, Horizontal scaling (using read replicas)
  • provides High Availability configuration provides data redundancy and failover capability with minimal downtime, when a zone or instance becomes unavailable due to a zonal outage, or an instance corruption
  • HA standby instance does not increase scalability and cannot be used for read queries.
  • Read replicas help scale horizontally the use of data in a database without degrading performance
  • is regional – although it now supports cross region read replicas
  • supports data encryption at rest and in transit
  • supports Point-In-Time recovery with binary logging and backups
  • available in two editions (introduced 2023-2024):
    • Cloud SQL Enterprise – provides all core capabilities, suitable for applications with less stringent availability and performance requirements
    • Cloud SQL Enterprise Plus – provides enhanced performance (up to 3x faster reads, 2x better write latency), higher availability, advanced observability (Query Insights), and supports Performance-optimized and Memory-optimized machine families (up to 32 GiB RAM per vCPU)
  • supports Private Service Connect (PSC) for automated endpoint creation in VPCs
  • supports Managed Connection Pooling with IAM authentication
  • supports PostgreSQL 18, MySQL 8.4, and SQL Server 2022
  • Extended Support – starting Feb 2025, instances running EOL major versions are automatically enrolled in paid extended support

AlloyDB for PostgreSQL (New Service)

  • fully managed, PostgreSQL-compatible database service designed for demanding enterprise workloads
  • more than 4x faster for transactional workloads and up to 100x faster for analytical queries compared to standard PostgreSQL
  • provides 99.99% availability SLA including maintenance
  • 100% compatible with open-source PostgreSQL
  • key features:
    • Columnar Engine – built-in columnar engine for real-time analytical queries on transactional data
    • AI/ML Integration – built-in vector search, Vertex AI integration for gen AI applications
    • Adaptive Autovacuum and automatic memory management
    • Index Advisor – recommends indexes to improve query performance
  • AlloyDB Omni – downloadable version that can run anywhere (on-premises, other clouds, edge)
  • supports PostgreSQL 14, 15, 16, 17, and 18
  • is regional with cross-region replication for disaster recovery

Cloud Spanner

  • fully managed, globally distributed, strongly consistent relational database service
  • provides virtually unlimited horizontal scalability with 99.999% availability SLA
  • supports SQL (GoogleSQL and PostgreSQL interface)
  • now available in editions (introduced 2024):
    • Standard Edition – core relational database capabilities
    • Enterprise Edition – adds multi-model capabilities including Spanner Graph, full-text search, vector search, managed autoscaling, and incremental backups
    • Enterprise Plus Edition – highest performance and availability
  • Multi-model capabilities (2024-2025):
    • Spanner Graph – native graph support using industry-standard GQL (Graph Query Language), interoperable with SQL for querying relational and connected data in a single operation
    • Vector Search – built-in vector similarity search for gen AI applications (cosine, Euclidean, dot-product distance)
    • Full-text Search – integrated text search capabilities
    • Vertex AI Integration – direct integration for embedding generation and AI model invocation
  • Tiered Storage (GA) – store data across SSDs or HDDs to optimize costs
  • is Global – spans multiple regions with strong consistency

Firestore (formerly Datastore)

  • Cloud Datastore has been superseded by Firestore and is now available in two modes:
    • Firestore in Datastore mode – same data model as original Datastore but runs on the Firestore engine (built on Spanner); backward-compatible with existing Datastore applications
    • Firestore in Native mode – new data model with real-time listeners, offline support, and richer querying
  • serverless, fully managed NoSQL document database
  • scales automatically with strong consistency
  • Firestore in Datastore mode improvements:
    • Queries in transactions are no longer required to be ancestor queries
    • Transactions are no longer limited to 25 entity groups
  • Firestore Enterprise Edition (2026) supports Text Search and Geospatial Search
  • Query Engine with Pipelines (2025) – introduces 100+ new pipeline operations for complex queries directly within the database
  • supports vector search for gen AI applications

BigQuery

  • serverless, highly scalable enterprise data warehouse for analytics
  • user- or project-level custom query quota
  • supports dry-run which helps in pricing queries based on the amount of bytes read i.e. --dry_run flag in the bq command-line tool or dryRun parameter when submitting a query job using the API
  • Pricing models:
    • On-demand – pay per TiB of data processed (first 1 TiB/month free)
    • Capacity (Editions) – replaced legacy flat-rate pricing (July 2023). Available in Standard, Enterprise, and Enterprise Plus editions with autoscaling slots
    • Legacy flat-rate and Flex Slots are no longer available for new purchases; existing commitments migrate to Editions upon expiration
  • BigQuery ML – build and run ML models using SQL, including:
    • TimesFM – state-of-the-art pre-trained forecasting model from Google Research
    • Gemini and open-source LLM model integration
    • Row-wise inference functions for mixing gen AI with SQL
    • Contribution Analysis for explaining changes in metrics
  • AI/ML capabilities (2024-2025):
    • AI functions for processing unstructured data
    • MCP (Model Context Protocol) support for agent-building tools
    • Over 100x scalability gains for LLM inference
    • Structured data generation/extraction with LLMs
  • BigQuery Studio – unified interface for data analytics with improved resource search and explorer

Bigtable

  • fully managed, wide-column NoSQL database designed for large analytical and operational workloads
  • handles massive scale with consistent low-latency (single-digit millisecond)
  • ideal for time-series data, IoT, financial data, and ad-tech
  • now available in editions (GA April 2026):
    • Enterprise and Enterprise Plus editions with advanced analytics, performance, and resource management features
  • GoogleSQL support – query Bigtable using standard SQL with features like:
    • Window functions for advanced analytic operations (GA 2026)
    • Distributed counting for real-time dashboards
    • KNN similarity search
  • Data Boost – serverless compute for running analytical queries without impacting operational workloads
  • In-Memory Tier (2026) – supports up to 120,000 QPS on a single row with hotspot resistance
  • supports replication across multiple regions for high availability
  • is regional with multi-region replication options

Filestore

  • fully managed, high-performance NFS file storage service
  • provides shared file storage mountable by Compute Engine VMs, GKE nodes, and other Google Cloud compute
  • supports NFSv3 and NFSv4.1 protocols
  • available in multiple tiers:
    • Basic (HDD/SSD) – for file sharing, software development, web hosting
    • Zonal – high-performance tier with higher IOPS and throughput
    • Enterprise – multi-zone with 99.99% availability SLA for business-critical apps
  • supports integration with GKE via Filestore CSI driver

Memorystore

  • fully managed in-memory data store service for Redis, Memcached, and Valkey
  • Memorystore for Valkey (GA 2025) – open-source, Redis-compatible in-memory database
    • 99.99% availability SLA
    • Supports Valkey versions 7.2, 8.0, and 9.0
    • Features: Private Service Connect, multi-VPC access, cross-region replication, persistence
    • Zero-downtime scaling, instances up to 14.5 TB
    • Valkey 9.0 includes SIMD optimizations for improved throughput and latency
  • Memorystore for Redis – managed Redis with Basic and Standard tiers (Standard includes replication and automatic failover)
  • Memorystore for Redis Cluster – high-throughput with clustering support
  • Memorystore for Memcached – managed Memcached for caching
  • supports vector search capabilities for gen AI applications

Google Persistent Disk & Hyperdisk

  • Persistent Disk – durable block storage for Compute Engine VMs
    • Standard (pd-standard) – HDD-backed, suitable for sequential read/write workloads
    • Balanced (pd-balanced) – SSD-backed, balance of performance and cost
    • SSD (pd-ssd) – SSD-backed, high random IOPS
    • Available as zonal or regional (synchronous replication across 2 zones)
  • Hyperdisk (newer generation, recommended) – higher performance block storage leveraging Google’s Titanium offload technology
    • Hyperdisk Balanced – general-purpose with configurable IOPS and throughput
    • Hyperdisk Extreme – highest IOPS for demanding databases
    • Hyperdisk Throughput – high throughput for bandwidth-intensive workloads at cost similar to cold storage
    • Hyperdisk ML – optimized for serving ML models with high throughput reads
    • Hyperdisk Balanced High Availability – for GKE workloads requiring HA (GKE 1.33+)
  • Hyperdisk Storage Pools (2024) – provision IOPS and throughput in aggregate; dynamically allocated across volumes for better utilization
  • Backup Vaults (GA 2025) – support for standalone Persistent Disk and Hyperdisk backups with multi-region capability

Google Local SSD

  • physically attached storage providing very high IOPS and low latency
  • ephemeral – data does not persist beyond the life of the instance
  • ideal for caches, scratch disks, and temporary processing
  • Titanium Local SSD (2024-2025) – next-generation local storage available with newer machine series:
    • C3 machine series with -lssd machine types (e.g., c3-standard-88-lssd)
    • C4 machine series with Intel Xeon 6 – up to 35% lower access latency
    • C4A with Google Axion processors – up to 72 vCPUs, 576 GB memory, 6 TB local storage
    • Z3 storage-optimized – up to 3x disk throughput, 35% lower latency, 3-36 TiB per VM
  • not available with shared-core machine types
  • data may not be available during maintenance events on storage-optimized VMs

Related Posts

Google Cloud SQL – Managed MySQL, PostgreSQL & SQL

~ Last updated on : ~ jayendrapatil

GCP Cloud SQL

  • Cloud SQL provides a cloud-based alternative to local MySQL, PostgreSQL, and Microsoft SQL Server databases
  • Cloud SQL is a managed solution that helps handle backups, replication, high availability and failover, data encryption, monitoring, and logging.
  • Cloud SQL is ideal for lift and shift migration from existing on-premises relational databases
  • Cloud SQL supports MySQL 5.6, 5.7, 8.0, 8.4, PostgreSQL (multiple versions), and SQL Server 2019, 2022, 2025

Cloud SQL Editions

  • Cloud SQL offers two editions: Cloud SQL Enterprise Plus and Cloud SQL Enterprise
  • Cloud SQL Enterprise Plus edition
    • Provides the best performance, availability, and observability for business-critical applications
    • Delivers up to 4x improved read performance using Data Cache (local SSD)
    • Delivers up to 3x higher write throughput and up to 98% lower write latency with Optimized Writes
    • Offers 99.99% availability SLA (inclusive of maintenance)
    • Provides near-zero downtime (<1 second) for planned maintenance and operations
    • Supports up to 128 vCPUs and 864 GB RAM (N2 machine series)
    • Supports Advanced Disaster Recovery with cross-region replication, switchover, and failover
    • Supports Read Pools with autoscaling for operational simplicity
    • Supports Managed Connection Pooling
    • Supports up to 35 days PITR log retention
    • Provides AI-assisted troubleshooting, enhanced Query Insights (30-day retention), and Index Advisor
    • Supports MySQL 8.0, 8.4 (MySQL 8.4 defaults to Enterprise Plus)
  • Cloud SQL Enterprise edition
    • Provides all core capabilities of Cloud SQL at a lower cost
    • Suitable for applications with less stringent availability and performance requirements
    • Offers 99.95% availability SLA (excludes maintenance)
    • Maintenance downtime of <60 seconds
    • Supports up to 7 days PITR log retention
    • Supports MySQL 5.6, 5.7, 8.0, 8.4
  • All existing Cloud SQL instances created before July 12, 2023 were automatically updated to Cloud SQL Enterprise edition
  • You can upgrade to Enterprise Plus edition using in-place upgrade with near-zero downtime

Cloud SQL High Availability

  • Cloud SQL instance HA configuration provides data redundancy and failover capability with minimal downtime, when a zone or instance becomes unavailable due to a zonal outage, or an instance corruption
  • HA configuration is also called a regional instance or cluster
  • With HA, the data continues to be available to client applications.
  • HA is made up of a primary and a standby instance and is located in a primary and secondary zone within the configured region
  • If an HA-configured instance becomes unresponsive, Cloud SQL automatically switches to serving data from the standby instance.
  • Data is synchronously replicated to each zone’s persistent disk, all writes made to the primary instance are replicated to disks in both zones before a transaction is reported as committed.
  • In the event of an instance or zone failure, the persistent disk is attached to the standby instance, and it becomes the new primary instance.
  • After a failover, the instance that received the failover continues to be the primary instance, even after the original instance comes back online.
  • Once the zone or instance that experienced an outage becomes available again, the original primary instance is destroyed and recreated and It becomes the new standby instance.
  • If a failover occurs in the future, the new primary will failover to the original instance in the original zone.
  • Cloud SQL Standby instance does not increase scalability and cannot be used for read queries
  • To see if failover has occurred, check the operation log’s failover history.
  • Write Endpoint (Enterprise Plus) – provides a DNS name that automatically resolves to the current primary instance IP, so applications don’t need to update connection strings after failover.

Cloud SQL High Availability

Cloud SQL Failover Process

  • Each second, the primary instance writes to a system database as a heartbeat signal.
  • Primary instance or zone fails.
  • If multiple heartbeats aren’t detected, failover is initiated. This occurs if the primary instance is unresponsive for approximately 60 seconds or the zone containing the primary instance experiences an outage.
  • Standby instance now serves data upon reconnection.
  • Through a shared static IP address with the primary instance, the standby instance now serves data from the secondary zone.
  • Users are then automatically rerouted to the new primary.

Cloud SQL Advanced Disaster Recovery

  • Advanced Disaster Recovery (DR) is available exclusively on Cloud SQL Enterprise Plus edition
  • Allows configuring cross-regional replication with a designated DR replica
  • Provides Replica Failover — promotes the DR replica immediately in the event of a regional failure
  • Provides Switchover — reverses the roles of the primary instance and a DR replica with zero data loss
  • Switchover can be used to restore a deployment to its original state after replica failover, or to test DR readiness
  • The DR replica is a cross-region read replica designated for disaster recovery
  • Uses a write endpoint to automatically redirect application traffic to the new primary after failover or switchover
  • Reduces RTO (Recovery Time Objective) significantly compared to manual promotion of cross-region replicas

Cloud SQL Read Replica

  • Read replicas help scale horizontally the use of data in a database without degrading performance
  • Read replica is an exact copy of the primary instance. Data and other changes on the primary instance are updated in almost real time on the read replica.
  • Read replica can be promoted if the original instance becomes corrupted.
  • Primary instance and read replicas all reside in Cloud SQL
  • Read replicas are read-only; you cannot write to them
  • Read replicas do not provide failover capability (use HA or Advanced DR instead)
  • Read replicas can now be configured with high availability for increased resilience
  • Google recommends limiting direct read replicas to 10 or fewer per primary instance. For additional replicas, use cascading read replicas.
  • During a zonal outage, traffic to read replicas in that zone stops.
  • Once the zone becomes available again, any read replicas in the zone will resume replication from the primary instance.
  • If read replicas are in a zone that is not in an outage, they are connected to the standby instance when it becomes the primary instance.
  • GCP recommends putting read replicas in a different zone from the primary and standby instances. for e.g., if you have a primary instance in zone A and a standby instance in zone B, put the read replicas in zone C. This practice ensures that read replicas continue to operate even if the zone for the primary instance goes down.
  • Client application needs to be configured to send reads to the primary instance when read replicas are unavailable.
  • Cloud SQL supports Cross-region replication that lets you create a read replica in a different region from the primary instance.
  • Cloud SQL supports External read replicas that are external MySQL instances which replicate from a Cloud SQL primary instance
  • Read replicas can have different vCPUs and memory from the primary instance but must have at least as much storage capacity.

Cascading Read Replicas

  • Cascading replication lets you create a read replica under another read replica in the same or a different region
  • Supports up to 4 levels of replicas in the hierarchy (including the primary instance)
  • A cascading replica can have up to 8 siblings (replicas from the same parent)
  • Use cases:
    • Disaster Recovery — cascading replicas in another region retain their own replicas when promoted
    • Performance — offloads replication work from the primary instance
    • Cost Reduction — only one cross-region replication incurs network egress; sub-replicas use free in-region transfer
    • Scale Reads — more replicas to share read load without burdening the primary
  • When a cascading replica is promoted, all its sub-replicas continue to replicate from it
  • You cannot delete a replica that has replicas under it; must start with leaf replicas

Read Pools (Enterprise Plus)

  • Read Pools provide a simplified, fully managed way to scale reads using multiple read replicas behind a single read endpoint
  • Available exclusively on Cloud SQL Enterprise Plus edition for MySQL and PostgreSQL
  • A read pool contains between 1 and 20 read pool nodes
  • Provides a single load balancer (read endpoint) that dispatches queries to nodes in round-robin fashion
  • Supports autoscaling — automatically adds or removes read pool nodes based on workload
  • You can add and remove replicas without making application changes
  • Simplifies connection management for read-heavy applications

Cloud SQL Point In Time Recovery

  • Point-in-time recovery (PITR) helps recover a Cloud SQL instance to a specific point in time
  • PITR uses write-ahead logs (for PostgreSQL) or binary logs (for MySQL)
  • PITR requires backups to be enabled for the instance
  • Point-in-time recovery is enabled by default when a new Cloud SQL instance is created
  • Log retention:
    • Cloud SQL Enterprise Plus edition: up to 35 days
    • Cloud SQL Enterprise edition: up to 7 days
  • Transaction logs are stored in the same region as the instance at no additional cost
  • PITR logs are stored in Cloud Storage (no longer on instance storage), eliminating storage impact on the instance

Cloud SQL Auth Proxy

  • Cloud SQL Auth Proxy (formerly known as Cloud SQL Proxy) provides secure access to instances without the need for Authorized networks or for configuring SSL.
    • Secure connections: Automatically encrypts traffic to and from the database using TLS 1.3 with a 256-bit AES cipher; SSL certificates are used to verify client and server identities.
    • Easier connection management: Handles authentication via IAM, removing the need to provide static IP addresses or manage SSL certificates.
    • IAM-based authorization: Uses IAM permissions to control which identities can connect to an instance.
  • Cloud SQL Auth Proxy does not provide a new connectivity path; it relies on existing IP connectivity. To connect to a Cloud SQL instance using private IP, the Cloud SQL Auth Proxy must be on a resource with access to the same VPC network as the instance.
  • Cloud SQL Auth Proxy works by having a local client running in the local environment. The application communicates with the Cloud SQL Auth Proxy with the standard database protocol used by the database.
  • Cloud SQL Auth Proxy uses a secure tunnel to communicate with its companion process running on the server.
  • While the proxy can listen on any port, it only creates outgoing connections to the Cloud SQL instance on port 3307.
  • For GKE deployments, the recommended pattern is running the Auth Proxy as a sidecar container in the same pod as the application.

Cloud SQL Auth Proxy

Cloud SQL Connectivity Options

  • Public IP — connect over the internet with authorized networks or Cloud SQL Auth Proxy
  • Private IP (Private Services Access) — connect using an internal IP address via VPC peering
  • Private Service Connect (PSC) — connect to Cloud SQL from multiple VPC networks across different projects, teams, or organizations without VPC peering
    • PSC provides a service attachment endpoint with a dedicated private IP
    • Works with both primary instances and read replicas
    • Can be combined with Private Services Access on the same instance
    • PSC Automation (Preview) simplifies deployment of PSC endpoints at scale
  • Cloud SQL Auth Proxy — IAM-authenticated, encrypted connections without managing SSL certificates or IP allowlists
  • Cloud SQL Language Connectors — open-source libraries for Java, Python, Go, and Node.js for simplified and secure connectivity
  • Managed Connection Pooling (Enterprise Plus) — built-in connection pooling to optimize database connection management

Cloud SQL AI and Vector Search

  • Cloud SQL integrates with Vertex AI to bring AI capabilities directly to your database
  • Supports generating vector embeddings using simple SQL functions (no external pipeline needed)
  • Supports vector storage and similarity search for MySQL and PostgreSQL, enabling gen AI use cases without a specialized vector database
  • Can invoke Vertex AI models (including Gemini) directly from SQL for online predictions
  • Supports building LLM-powered applications using LangChain integration
  • Supports Model Endpoint Management — register, invoke, and manage AI models from within Cloud SQL
  • Enables RAG (Retrieval-Augmented Generation) workflows with vector search capabilities

Cloud SQL Features Comparison

Cloud SQL Features Comparison

GCP Certification Exam Practice Questions

  • Questions are collected from Internet and the answers are marked as per my knowledge and understanding (which might differ with yours).
  • GCP services are updated everyday and both the answers and questions might be outdated soon, so research accordingly.
  • GCP exam questions are not updated to keep up the pace with GCP updates, so even if the underlying feature has changed the question might not be updated
  • Open to further feedback, discussion and correction.
  1. You work for a mid-sized enterprise that needs to move its operational system transaction data from an on-premises database to GCP. The database is about 20 TB in size. Which database should you choose?
    1. Cloud SQL
    2. Cloud Bigtable
    3. Cloud Spanner
    4. Cloud Datastore
  2. An application that relies on Cloud SQL to read infrequently changing data is predicted to grow dramatically. How can you increase capacity for more read-only clients?
    1. Configure high availability on the master node
    2. Establish an external replica in the customer’s data center
    3. Use backups so you can restore if there’s an outage
    4. Configure read replicas.
  3. A Company is using Cloud SQL to host critical data. They want to enable high availability in case a complete zone goes down. How should you configure the same?
    1. Create a Read replica in the same region different zone
    2. Create a Read replica in the different region different zone
    3. Create a Failover replica in the same region different zone
    4. Create a Failover replica in the different region different zone
  4. A Company is using Cloud SQL to host critical data. They want to enable Point In Time recovery (PIT) to be able to recover the instance to a specific point in time. How should you configure the same?
    1. Create a Read replica for the instance
    2. Switch to Spanner 3 node cluster
    3. Create a Failover replica for the instance
    4. Enable Binary logging and backups for the instance
  5. A company needs a Cloud SQL deployment that provides 99.99% availability SLA inclusive of maintenance window downtime. Which configuration should they choose?
    1. Cloud SQL Enterprise edition with HA enabled
    2. Cloud SQL Enterprise edition with read replicas in multiple zones
    3. Cloud SQL Enterprise Plus edition with HA enabled
    4. Cloud SQL Enterprise Plus edition with read replicas only
  6. A company wants to set up cross-region disaster recovery for their Cloud SQL database with the ability to perform switchover drills with zero data loss. What should they use?
    1. Cross-region read replicas with manual promotion
    2. Cloud SQL Enterprise Plus edition with Advanced Disaster Recovery
    3. Cloud SQL Enterprise edition with automated backups in another region
    4. Cloud Spanner with multi-region configuration
  7. A company wants to scale read traffic for their Cloud SQL MySQL database by adding multiple read replicas that can be accessed via a single endpoint and auto-scaled based on demand. What feature should they use?
    1. Cross-region read replicas with DNS load balancing
    2. Cascading read replicas in the same region
    3. Cloud SQL Read Pools (Enterprise Plus)
    4. External read replicas with a custom load balancer
  8. An organization wants to connect to a Cloud SQL instance from multiple VPCs across different projects without using VPC peering. Which connectivity option should they choose?
    1. Cloud SQL Auth Proxy with public IP
    2. Private Services Access with shared VPC
    3. Private Service Connect (PSC)
    4. Authorized networks with IP allowlisting

See also: Google Cloud Storage Services Cheat Sheet

References

 

Google Cloud Storage & Database Options Comparison

~ Last updated on : ~ jayendrapatil ~ 2 Comments

GCP Storage Options

GCP provides various storage options and the selection can be based on

  • Structured vs Unstructured
  • Relational (SQL) vs Non-Relational (NoSQL)
  • Transactional (OLTP) vs Analytical (OLAP)
  • Fully Managed vs Requires Provisioning
  • Global vs Regional
  • Horizontal vs Vertical scaling

Cloud Firestore

  • Cloud Firestore is a fully managed, highly scalable, serverless, non-relational NoSQL document database
  • fully managed with no-ops and no planned downtime and no need to provision database instances (vs Bigtable)
  • uses a distributed architecture to automatically manage scaling.
  • queries scale with the size of the result set, not the size of the data set
  • supports ACID Atomic transactions – all or nothing (vs Bigtable)
  • provides High availability of reads and writes – runs in Google data centers, which use redundancy to minimize impact from points of failure.
  • provides massive scalability with high performance – uses a distributed architecture to automatically manage scaling.
  • scales from zero to terabytes with flexible storage and querying of data
  • provides SQL-like query language
  • supports strong consistency
  • supports data encryption at rest and in transit
  • provides terabytes of capacity with a maximum unit size of 1 MB per entity (vs Bigtable)
  • Firestore Editions (2025)
    • Standard edition – core Firestore capabilities with standard querying support
    • Enterprise edition – provides MongoDB API compatibility, a new pipeline query engine with 200+ query operations, additional data types, new index types, and text/geospatial search
  • Enterprise Edition Features
    • MongoDB Compatibility (GA Aug 2025) – use existing MongoDB application code, drivers, and tools as a drop-in replacement while getting Firestore’s auto-scaling and high availability
    • Pipeline Query Engine – supports 200+ new query capabilities (pipeline operations) for complex queries directly within the database
    • Text Search and Geospatial Search – native full-text and geospatial query support without external services
    • Maximum document size increased to 16 MiB (Enterprise edition)
    • Indexes are not required for queries in Enterprise edition
  • Consider using Cloud Firestore if you need to store semi-structured objects, or if require support for transactions and SQL-like queries.

Cloud Bigtable

  • Bigtable provides a scalable, fully managed, non-relational NoSQL wide-column analytical big data database service suitable for both low-latency single-point lookups and precalculated analytics.
  • supports large quantities (>1 TB) of semi-structured or structured data (vs Datastore)
  • supports high throughput or rapidly changing data (vs BigQuery)
  • managed, but needs provisioning of nodes and can be expensive (vs Datastore and BigQuery)
  • does not support transactions or strong relational semantics (vs Datastore)
  • Now supports GoogleSQL queries (GA 2024) – familiar SQL syntax for querying Bigtable data directly
  • Not Transactional and does not support ACID
  • provides eventual consistency
  • ideal for time-series or natural semantic ordering data
  • can run asynchronous batch or real-time processing on the data
  • can run machine learning algorithms on the data
  • provides petabytes of capacity with a maximum unit size of 10 MB per cell and 100 MB per row.
  • Bigtable Editions (GA April 2026)
    • Enterprise edition – advanced features in performance, analytic query capability, and resource management
    • Enterprise Plus edition – includes in-memory tier with sub-millisecond latency and hotspot resistance supporting up to 120,000 queries per second on a single row
  • New Features (2024-2026)
    • Bigtable SQL (GoogleSQL) – query data using familiar SQL syntax with specialized features preserving flexible schema
    • Data Boost – serverless analytical queries without impacting operational workloads
    • Incremental Materialized Views – simplify creation of real-time metrics
    • Window Functions (GA April 2026) – advanced analytic operations over multiple table rows
    • KNN Vector Search – K nearest neighbors similarity search for AI/ML use cases
    • Distributed Counting – instant metric retrieval for real-time dashboards
    • In-Memory Tier – hotspot resistance with sub-millisecond latency
    • Agent Skills (April 2026) – let AI agents assist with schema design, SQL queries, and infrastructure management
  • Usage Patterns
    • Low-latency read/write access
    • High-throughput data processing
    • Time series support
  • Anti Patterns
    • Not an ideal storage option for future analysis – Use BigQuery instead
    • Not an ideal storage option for transactional data – Use relational database or Datastore
  • Common Use cases
    • IoT, finance, adtech
    • Personalization, recommendations
    • Monitoring
    • Geospatial datasets
    • Graphs
    • Real-time AI/ML inference and vector search
  • Consider using Cloud Bigtable, if you need high-performance datastore to perform analytics on a large number of structured objects

Cloud Storage

  • Cloud Storage provides durable and highly available object storage.
  • fully managed, simple administration, cost-effective, and scalable service that does not require capacity management
  • supports unstructured data storage like binary or raw objects
  • provides high performance, internet-scale
  • supports data encryption at rest and in transit
  • provides 99.999999999% (11 nines) annual durability
  • Storage Classes: Standard, Nearline (30-day min), Coldline (90-day min), Archive (365-day min)
  • Autoclass – automatically transitions objects to appropriate storage classes based on access patterns
  • New Features (2024-2026)
    • Cloud Storage Rapid (2025-2026) – high-performance storage tier for AI/ML workloads
      • Rapid Bucket (formerly Rapid Storage) – zonal object storage with <1ms random read/write latency, 6 TB/s throughput
      • Rapid Cache (formerly Anywhere Cache) – accelerates reads and colocates compute with data, up to 20 Tbps throughput
    • Smart Storage – automated metadata annotation for unstructured data with AI agent connectivity via MCP
    • Storage Intelligence – zero-configuration dashboards, aggregated activity views, and enhanced batch operations
    • Bucket Relocation – move buckets between regions with minimal downtime
    • Batch Operations Dry Run Mode – simulate batch jobs without modifying data
  • Consider using Cloud Storage, if you need to store immutable blobs larger than 10 MB, such as large images or movies. This storage service provides petabytes of capacity with a maximum unit size of 5 TB per object.
  • Usage Patterns
    • Images, pictures, and videos
    • Objects and blobs
    • Unstructured data
    • Long term storage for archival or compliance
    • AI/ML training data and model checkpoints
  • Anti Patterns
    • Not ideal for structured/relational data
    • Not ideal for frequently changing data requiring low-latency updates
  • Common Use cases
    • Storing and streaming multimedia
    • Storage for custom data analytics pipelines
    • Archive, backup, and disaster recovery
    • AI/ML training datasets and model serving

Cloud SQL

  • provides fully managed, relational SQL databases
  • offers MySQL, PostgreSQL, and SQL Server databases as a service
  • manages OS & Software installation, patches and updates, backups and configuring replications, failover however needs to select and provision machines (vs Cloud Spanner)
  • single region only – although it now supports cross-region read replicas (vs Cloud Spanner)
  • Cloud SQL Editions
    • Enterprise edition – core capabilities, suitable for applications with less stringent availability/performance requirements. Up to 96 vCPU, 624 GB RAM.
    • Enterprise Plus edition – highest performance with optimized software/hardware stack. Up to 128 vCPU, 864 GB RAM. Includes data cache, up to 35-day point-in-time log retention, sub-second maintenance downtime, and advanced disaster recovery.
  • Scaling
    • provides vertical scalability (Max. storage of 64 TB)
    • storage can be increased without incurring any downtime
    • provides an option to increase the storage automatically
    • storage CANNOT be decreased
    • supports Horizontal scaling for read-only using read replicas (vs Cloud Spanner)
    • performance is linked to the disk size
  • Security
    • data is encrypted when stored in database tables, temporary files, and backups.
    • external connections can be encrypted by using SSL, or by using the Cloud SQL Proxy.
    • Private Service Connect (PSC) support for simplified private connectivity
  • High Availability
    • fault-tolerance across zones can be achieved by configuring the instance for high availability by adding a failover replica
    • failover is automatic
    • can be created from primary instance only
    • replication from the primary instance to failover replica is semi-synchronous.
    • failover replica must be in the same region as the primary instance, but in a different zone
    • only one instance for every primary instance allowed
    • supports managed backups and backups are created on primary instance only
    • supports automatic replication
    • Enterprise Plus: sub-second maintenance downtime (vs up to 120 seconds for Enterprise)
  • Backups
    • Automated backups can be configured and are stored for 7 days
    • Manual backups (snapshots) can be created and are not deleted automatically
    • Fast Clone (GA) – clone operations within the same zone for rapid environment creation
  • Point-in-time recovery
    • requires binary logging enabled.
    • every update to the database is written to an independent log, which involves a small reduction in write performance.
    • performance of the read operations is unaffected by binary logging, regardless of the size of the binary log files.
    • Enterprise Plus: up to 35-day log retention (vs 7 days for Enterprise)
  • Usage Patterns
    • direct lift and shift for MySQL, PostgreSQL, SQL Server database only
    • relational database service with strong consistency
    • OLTP workloads
  • Anti Patterns
    • need data storage more than 64 TB or horizontal write scaling, use Cloud Spanner
    • need global availability with low latency, use Cloud Spanner
    • not a direct replacement for Oracle – use installation on GCE or consider AlloyDB for PostgreSQL workloads
  • Common Use cases
    • Websites, blogs, and content management systems (CMS)
    • Business intelligence (BI) applications
    • ERP, CRM, and eCommerce applications
    • Geospatial applications
  • Consider using Cloud SQL for full relational SQL support for OLTP and lift and shift of MySQL, PostgreSQL, SQL Server databases

Cloud Spanner

  • Cloud Spanner provides fully managed, relational SQL databases with joins and secondary indexes
  • provides cross-region, global, horizontal scalability, and availability
  • supports strong consistency, including strongly consistent secondary indexes
  • provides high availability through synchronous and built-in data replication.
  • provides strong global consistency
  • supports database sizes exceeding ~2 TB (vs Cloud SQL)
  • does not provide direct lift and shift for relational databases (vs Cloud SQL)
  • expensive as compared to Cloud SQL
  • Multi-Model Database (2024-2025)
    • Spanner Graph (GA Jan 2025) – supports industry-standard Graph Query Language (GQL) with full SQL interoperability for querying structured and connected data
    • Vector Search – native vector embeddings and similarity search for AI/ML and RAG applications
    • Full-Text Search – native text search capabilities without external services
    • Hybrid Search – combine vector search, full-text search, and ML model reranking in a unified platform
    • Vertex AI Integration – native integration for model serving and inferencing with SQL
  • Spanner Omni (2026 Preview)
    • Self-managed version of Spanner that runs on-premises, across clouds, or on a laptop
    • Brings Spanner’s scalability, high availability, strong consistency, and multi-model capabilities anywhere
    • Supports air-gapped or connected deployments, single machine to clusters of thousands
  • Tiered Storage (GA) – store data across SSD or HDD tiers for cost optimization
  • Consider using Cloud Spanner for full relational SQL support, with horizontal scalability spanning petabytes for OLTP, or as a multi-model database supporting relational, graph, vector, and text search workloads

BigQuery

  • provides fully managed, no-ops, OLAP, enterprise data warehouse (EDW) with SQL and fast ad-hoc queries.
  • provides high capacity, data warehousing analytics solution
  • ideal for big data exploration and processing
  • not ideal for operational or transactional databases
  • provides SQL interface
  • A scalable, fully managed data-to-AI platform
  • BigQuery Editions – Standard, Enterprise, and Enterprise Plus with different pricing and feature tiers
  • New Features (2024-2026)
    • Conversational Analytics (Preview Jan 2026) – analyze data using natural language with AI-powered data agents that understand context and generate SQL
    • BigQuery Graph – uncover complex relationships and patterns in data
    • Vector Search – embeddings and hybrid search for RAG applications
    • BigQuery ML – train and run ML models directly in BigQuery using SQL
    • Data Engineering Agent – automates data preparation, error detection, and pipeline building
    • Data Science Agent – automates data loading, feature engineering, model training and evaluation
    • BigQuery Studio – unified workspace with Gemini-powered assistant for resource discovery and query generation
    • MCP Integration – Model Context Protocol for AI agent connectivity
  • Usage Patterns
    • OLAP workloads up to petabyte-scale
    • Big data exploration and processing
    • Reporting via business intelligence (BI) tools
    • AI/ML model training and inference at scale
  • Anti Patterns
    • Not an ideal storage option for transactional data or OLTP – Use Cloud SQL or Cloud Spanner instead
    • Low-latency read/write access – Use Bigtable instead
  • Common Use cases
    • Analytical reporting on large data
    • Data science and advanced analyses
    • Big data processing using SQL
    • GenAI and agentic AI applications with data

AlloyDB for PostgreSQL

  • AlloyDB is a fully managed, PostgreSQL-compatible database designed for enterprise-grade OLTP and hybrid transactional/analytical (HTAP) workloads
  • wire-compatible with PostgreSQL 14 and 15 – existing drivers, ORMs, and most extensions work without modification
  • provides up to 4x faster for transactional workloads and up to 100x faster for analytical queries compared to standard PostgreSQL
  • uses a scale-out architecture with compute and storage separation
  • built-in AI capabilities with Google’s cutting-edge technology
  • AlloyDB AI
    • Generate vector embeddings from within the database
    • Native vector search with up to 10x faster index creation and 4x faster search queries
    • Filtered vector search up to 10x faster than standard PostgreSQL HNSW
    • Integration with Vertex AI for model serving and inferencing
    • AlloyDB AI query engine with Vertex AI Ranking API
  • AlloyDB Omni – downloadable version that runs on-premises or in other clouds
  • 99.99% availability SLA with automated backups, replication, and failover
  • Usage Patterns
    • Enterprise PostgreSQL workloads requiring high performance
    • HTAP (hybrid transactional/analytical) workloads
    • AI-powered applications requiring vector search
    • Migration from commercial databases (Oracle, SQL Server) to PostgreSQL
  • Anti Patterns
    • Need global horizontal scaling – Use Cloud Spanner
    • Need non-relational/NoSQL – Use Firestore or Bigtable
    • Need MySQL or SQL Server compatibility – Use Cloud SQL
  • Consider using AlloyDB for PostgreSQL workloads requiring high performance, AI integration, or migration from commercial databases

Memorystore

  • provides scalable, secure, and highly available in-memory service
  • fully managed as provisioning, replication, failover, and patching are all automated
  • is protected from the internet using VPC networks and private IP and comes with IAM integration
  • Supported Engines
    • Memorystore for Valkey (GA 2025) – open-source, high-performance key-value store (successor to Redis OSS). Supports Valkey 8.0 and 9.0. 99.99% availability SLA, instances up to 14.5 TB, cross-region replication, Private Service Connect, multi-VPC access.
    • Memorystore for Redis Cluster – managed Redis cluster mode with zero-downtime scaling
    • Memorystore for Redis – standard Redis instances (standalone and high availability)
    • Memorystore for Memcached – distributed in-memory caching
  • Valkey 9.0 Features (GA 2026)
    • SIMD optimizations for improved throughput and latency
    • Enhanced performance over previous versions
    • Full compatibility with Redis OSS commands
  • Usage Patterns
    • Lift and shift migration of applications
    • Low latency data caching and retrieval
    • Session management
    • Real-time leaderboards and counting
  • Anti Patterns
    • Relational or NoSQL database
    • Analytics solution
    • Persistent primary data store (use as cache layer)
  • Common Use cases
    • User session management
    • Application caching
    • Real-time analytics and pub/sub
    • Gaming leaderboards

GCP Storage Options Decision Tree

GCP Storage Options Decision Tree

GCP Certification Exam Practice Questions

  • Questions are collected from Internet and the answers are marked as per my knowledge and understanding (which might differ with yours).
  • GCP services are updated everyday and both the answers and questions might be outdated soon, so research accordingly.
  • GCP exam questions are not updated to keep up the pace with GCP updates, so even if the underlying feature has changed the question might not be updated
  • Open to further feedback, discussion and correction.
  1. Your application is hosted across multiple regions and consists of both relational database data and static images. Your database has over 10 TB of data. You want to use a single storage repository for each data type across all regions. Which two products would you choose for this task? (Choose two)
    1. Cloud Bigtable
    2. Cloud Spanner
    3. Cloud SQL
    4. Cloud Storage
  2. You are building an application that stores relational data from users. Users across the globe will use this application. Your CTO is concerned about the scaling requirements because the size of the user base is unknown. You need to implement a database solution that can scale with your user growth with minimum configuration changes. Which storage solution should you use?
    1. Cloud SQL
    2. Cloud Spanner
    3. Cloud Firestore
    4. Cloud Datastore
  3. Your company processes high volumes of IoT data that are time-stamped. The total data volume can be several petabytes. The data needs to be written and changed at a high speed. You want to use the most performant storage option for your data. Which product should you use?
    1. Cloud Datastore
    2. Cloud Storage
    3. Cloud Bigtable
    4. BigQuery
  4. Your App Engine application needs to store stateful data in a proper storage service. Your data is non-relational database data. You do not expect the database size to grow beyond 10 GB and you need to have the ability to scale down to zero to avoid unnecessary costs. Which storage service should you use?
    1. Cloud Bigtable
    2. Cloud Dataproc
    3. Cloud SQL
    4. Cloud Firestore (Datastore mode)
  5. A financial organization wishes to develop a global application to store transactions happening from different part of the world. The storage system must provide low latency transaction support and horizontal scaling. Which GCP service is appropriate for this use case?
    1. Bigtable
    2. Datastore
    3. Cloud Storage
    4. Cloud Spanner
  6. You work for a mid-sized enterprise that needs to move its operational system transaction data from an on-premises database to GCP. The database is about 20 TB in size. Which database should you choose?
    1. Cloud SQL
    2. Cloud Bigtable
    3. Cloud Spanner
    4. Cloud Datastore

    Note: With Cloud SQL now supporting up to 64 TB, Cloud SQL could also be a valid option for 20 TB. However, for operational transactional data requiring high scalability, Cloud Spanner remains the better choice.

  7. Your team needs a PostgreSQL-compatible database that can handle both transactional and analytical queries with high performance. The application also requires built-in vector search capabilities for an AI-powered recommendation engine. Which GCP service should you choose?
    1. Cloud SQL for PostgreSQL
    2. AlloyDB for PostgreSQL
    3. Cloud Spanner
    4. BigQuery
  8. Your company is building a real-time fraud detection system that needs to query relationships between entities (accounts, transactions, merchants) while also performing vector similarity searches on transaction patterns. The system must provide strong consistency and global availability. Which database should you use?
    1. Cloud Bigtable
    2. BigQuery
    3. Cloud Spanner
    4. Cloud Firestore
  9. Your organization is migrating from MongoDB to Google Cloud. You want to minimize code changes and use existing MongoDB drivers and tools. The application requires automatic scaling and high availability. Which GCP service should you use?
    1. Cloud SQL for PostgreSQL
    2. Cloud Bigtable
    3. Cloud Firestore (Enterprise edition with MongoDB compatibility)
    4. AlloyDB for PostgreSQL
  10. You need a high-performance caching layer for your microservices application on GCP. The solution must support cross-region replication, provide 99.99% availability, and be compatible with open-source tooling. Which service should you choose?
    1. Memorystore for Redis
    2. Cloud CDN
    3. Memorystore for Valkey
    4. Cloud Firestore

See also: Google Cloud Storage Services Cheat Sheet