Prepare and test your skills

Prepare and test your skills

Storage models in Google Cloud divide data work by data structure, latency expectations, scalability demands, and durability requirements. Cloud Storage (opens in a new tab) provides object storage, Cloud SQL (opens in a new tab) and Cloud Spanner (opens in a new tab) provide relational storage, Firestore (opens in a new tab) and Cloud Bigtable (opens in a new tab) provide NoSQL storage, and BigQuery provides analytical storage. Choosing among them depends on functional characteristics, resilience requirements, performance expectations, and cost goals.
Object storage with Cloud Storage holds unstructured data as objects in buckets at low cost with strong durability. The bucket location controls how data is replicated and how well it survives zone or region outages. Because replication is built into the service, durability comes from placement rather than from managing copies yourself.
Bucket location determines replication behavior and recovery characteristics:
Synchronous replication within a region supports zero or near-zero recovery time objective (RTO) and recovery point objective (RPO). RTO is how fast data is available again, and RPO is how much recent data could be lost. For dual-region buckets, turbo replication ensures objects are replicated across the region pair with an RPO of 15 minutes.
Relational storage with Cloud SQL and Cloud Spanner organizes data into tables with rows, columns, and defined relationships. This model fits operational workloads that need consistent transactions and structured queries. Both services are managed services that provide high availability (HA) backed by service level agreements (SLAs) with built-in backup mechanisms and resilience features.
Cloud Spanner is a managed multi-regional service that supports an active-active topology across two regions. Data flows from the active workload to replicas in multiple locations, so the service remains available when one zone or region has a problem. Many managed services provide synchronous replication between zones within a single region by default, which supports zero or near-zero RTO and RPO without a single point of failure. Choose relational storage when correctness of related records matters more than flexible structure or massive object scale.
NoSQL storage with Firestore and Cloud Bigtable stores operational data in flexible, non-relational structures instead of fixed tables. This model fits workloads with changing fields, very large numbers of records, and high scalability demands. The application writes flexible records to the database, and the service distributes the data to handle growth and maintain availability.
Firestore and Cloud Bigtable are managed services with HA backed by SLAs and built-in resilience. Like other managed data services, they reduce operational work for backups, monitoring, availability, and recovery compared with running a third-party database on virtual machines. Choose NoSQL storage when the workload needs flexible structure and elastic scale for operational reads and writes rather than strict relational schema or large-scale historical analysis.
Analytical storage with BigQuery runs queries over large datasets to find trends and support decisions. The analytical model separates large-scale analysis from day-to-day operational reads and writes. Data flows from operational systems and object storage into BigQuery, where it is retained for repeated querying.
BigQuery is a managed service, so HA, scalability, and resilience are built into the platform rather than managed server by server. Increasing availability generally costs more and can add operational complexity, so analytical workloads balance query performance, data durability, and cost. Choose analytical storage when the goal is to analyze large volumes of data rather than store individual objects or process single operational transactions.
Choosing between the four storage models means comparing structure, scale pattern, and durability approach. Object storage emphasizes low-cost replication of files across zones or regions. Relational storage emphasizes correct transactions with synchronous replication, including multi-regional active-active for Spanner. NoSQL emphasizes elastic scale for high-volume operations, while analytical storage emphasizes managed scale for repeated queries.
| Model | Example services | Data structure | Durability and scale emphasis |
|---|---|---|---|
| Object | Cloud Storage | Unstructured objects in buckets | Low-cost replication across zones or across regions |
| Relational | Cloud SQL, Cloud Spanner | Tables with rows and relationships | Synchronous replication with zero or near-zero RTO and RPO, including multi-regional active-active for Spanner |
| NoSQL | Firestore, Cloud Bigtable | Flexible non-relational records | Elastic scale for high-volume operational workloads with managed HA |
| Analytical | BigQuery | Large datasets organized for analysis | Managed scale and resilience for repeated queries over historical data |
Exam tip: Regional data is replicated synchronously across zones, while dual-region and multi-region metadata is written synchronously and data is replicated asynchronously, with turbo replication providing a 15-minute RPO for dual-region buckets.
Service capabilities and architectural fit describe how each Google Cloud data service matches a workload pattern. Cloud Storage (opens in a new tab) fits unstructured retention, Cloud SQL (opens in a new tab) fits regional transactions, Cloud Spanner (opens in a new tab) fits global transactions, Cloud Bigtable (opens in a new tab) fits massive operations, BigQuery fits large-scale analytics, and Firestore (opens in a new tab) fits app-facing documents. Choosing the right service means matching its design to whether the workload is transactional, operational, analytical, or focused on unstructured data retention.
Cloud Storage for unstructured data retention stores objects such as images, videos, backups, logs, and data lake files. Data is stored in buckets, and bucket location affects durability and availability. Regional buckets replicate data synchronously across zones in one region, while dual-region and multi-region buckets store data redundantly in at least two geographic locations. For dual-region buckets, turbo replication can replicate objects across region pairs with an RPO of 15 minutes. Use Cloud Storage when the workload needs cheap, durable retention or a staging area for analytics, not transactional queries.
Cloud SQL for regional transactional systems is a managed relational database service for MySQL, PostgreSQL, and SQL Server. It supports traditional transactional processing and operational querying with SQL, and it provides built-in backups, high availability, and scaling features. Regional high availability uses synchronous replication across zones within a region. For SQL Server workloads, Cloud SQL is recommended unless you need operating system access, in which case a failover cluster instance with NetApp Volumes may be used. Choose Cloud SQL for lift-and-shift applications, web apps, ERP, CRM, and line-of-business systems that need a regional relational database.
Cloud Spanner for globally consistent transactions is a globally distributed, horizontally scalable relational database. It provides global consistency and strong consistency across regions, making it a managed multi-regional service for critical replication. Spanner supports SQL and relational schemas while scaling horizontally for high transaction volumes. Data flows from the primary write location to replicas in other regions, so applications in different regions see the same correct data. Use Spanner for transactional processing at global scale, such as financial services, inventory, and gaming systems that need relational semantics plus global consistency.
Cloud Bigtable for massive operational throughput is a wide-column NoSQL database built for large analytical and operational workloads. It scales horizontally to handle high throughput and low latency, and it is a common target for HBase migrations. Bigtable fits time-series data, IoT telemetry, financial market data, and personalization workloads. It is not designed for complex joins or multi-row transactions, so choose it for operational querying at massive scale rather than traditional relational processing.
BigQuery for serverless analytics at scale is a serverless, highly scalable data warehouse for analytics. It uses columnar storage and SQL queries to support large-scale analytics, business intelligence, and data warehousing. The BigQuery Data Transfer Service automates data movement from software-as-a-service apps to BigQuery on a scheduled, managed basis. Data can also flow from Cloud Storage into BigQuery for analysis. Choose BigQuery for analytical processing over large datasets, not for transactional workloads.
Firestore for flexible app data with live sync is a document NoSQL database for mobile, web, and IoT applications. It offers schema flexibility, real-time synchronization, and offline support, so application data can change shape without rigid table migrations. Firestore scales automatically for app-facing workloads such as user profiles, catalogs, and real-time collaboration. Data flows from the app client to Firestore and then to other connected clients when data changes. It is not intended for large-scale analytics or high-throughput wide-column operational workloads.
How transactional, operational, and analytical services connect explains common architectures that combine the six services. Cloud Storage can hold raw or retained data that later flows into BigQuery for analytics. Cloud SQL and Cloud Spanner handle transactional systems, while Bigtable and Firestore serve operational application access patterns. BigQuery then supports analytical querying across large datasets without affecting the transactional databases.
The services differ by data type, architectural fit, and common business use case:
Exam tip: Choose Cloud Spanner when you need relational transactions with global consistency and horizontal scale; choose Cloud SQL when a regional managed relational database is enough.
Common business use cases for each data service show which organizational scenario fits each Google Cloud option. Cloud Storage (opens in a new tab) fits archiving, Cloud Spanner (opens in a new tab) fits global transactions, Cloud SQL (opens in a new tab) fits regional business systems, Cloud Bigtable (opens in a new tab) fits real-time operations, BigQuery fits reporting, and Firestore (opens in a new tab) fits mobile backends. The six services cover object storage, relational databases, NoSQL databases, and analytics warehouses.
Archiving and static content with Cloud Storage uses object storage for unstructured data such as files, images, videos, backups, and archives. It stores data as objects in buckets, and buckets can be regional, dual-region, or multi-region; dual-region and multi-region buckets store data redundantly in at least two geographic locations. Common business uses include archiving, backup and disaster recovery, serving static content, and acting as the target storage system for data migrations. Choose Cloud Storage when you need low-cost, durable, scalable storage without SQL queries or low-latency key-value access.
Global ecommerce and ledgers with Cloud Spanner use a globally distributed relational database that provides strong consistency and horizontal scaling. It stores structured data in tables and supports SQL and transactions. It is a managed multi-regional service, which makes it useful when data replication across regions is critical. Common business uses include global ecommerce transactions, financial ledgers, inventory systems, and gaming leaderboards. Choose Cloud Spanner when a relational workload must scale globally and stay highly available across regions.
Regional business applications with Cloud SQL use a managed relational database service for MySQL, PostgreSQL, and SQL Server. It stores structured data in tables and is designed for traditional SQL applications. Common business uses include ecommerce transactions, enterprise reporting, content management systems, ERP and CRM applications, and lift-and-shift migrations of existing databases. Choose Cloud SQL when you need a familiar relational database with ACID transactions and do not need global horizontal scale.
Real-time operations with Cloud Bigtable use a NoSQL wide-column database built for high-throughput, low-latency reads and writes at very large scale. It stores data in rows and columns with a row key, similar to an HBase table, and it is a target for HBase migrations. Common business uses include real-time analytics, IoT and time-series data, ad technology, personalization, and recommendation engines. Choose Bigtable when you need fast access to massive amounts of operational data and can design around row keys instead of complex SQL joins.
Reporting and machine learning with BigQuery use a serverless, highly scalable data warehouse for analytics. It stores structured and semi-structured data in tables and lets users run SQL queries over large datasets without managing servers. The BigQuery Data Transfer Service automates data movement from software-as-a-service apps to BigQuery on a scheduled, managed basis. Common business uses include enterprise reporting, business intelligence dashboards, real-time analytics, data lake exploration, and machine learning. Choose BigQuery when the goal is to analyze large amounts of data quickly and cost-effectively.
Mobile and collaborative apps with Firestore use a NoSQL document database for mobile, web, and IoT applications. Firestore, also called Cloud Firestore, stores data in documents organized into collections, and it supports real-time synchronization and offline access. Common business uses include mobile application backends, user profiles, chat apps, catalogs, and collaborative applications. Data flows from one client to Firestore and then to other connected clients for live updates. Choose Firestore when an application needs a flexible document model, automatic scaling, and live updates for connected clients.
Matching business scenarios to the best-fit service depends on the primary workload and access pattern. Archiving needs cheap retention, transactions need relational correctness, real-time operations need low-latency scale, mobile apps need live sync, and reporting needs analysis over large data. Cost, performance, and functional fit together determine the choice.
A decision tree that branches on workload type: unstructured data leads to Cloud Storage, analytics to BigQuery, relational transactions to Cloud SQL or Spanner, and NoSQL operational data to Firestore or Bigtable, each with its common business use case.
| Business scenario | Best-fit service | Why it fits |
|---|---|---|
| Archiving | Cloud Storage | Low-cost object storage for files, backups, and long-term retention. |
| Ecommerce transactions | Cloud SQL or Cloud Spanner | Cloud SQL handles regional relational transactions; Cloud Spanner handles global, strongly consistent relational transactions. |
| Real-time analytics | BigQuery or Cloud Bigtable | BigQuery analyzes large datasets with SQL; Bigtable serves high-throughput, low-latency operational data. |
| Mobile application backend | Firestore | Document model with real-time sync and offline support for mobile and web clients. |
| Enterprise reporting | BigQuery or Cloud SQL | BigQuery supports large-scale analytical reporting; Cloud SQL supports reporting from operational relational data. |
Use Cloud Spanner when the relational workload needs global scale with strong consistency across regions, such as global ecommerce transactions, financial ledgers, inventory systems, or gaming systems. Use Cloud SQL when a regional managed relational database for MySQL, PostgreSQL, and SQL Server is enough for web apps, ERP, CRM, and line-of-business systems with synchronous replication across zones within a region.
Cloud Bigtable is a wide-column NoSQL database for high-throughput, low-latency operational reads and writes at very large scale and does not support complex joins or multi-row transactions. BigQuery is a serverless data warehouse that uses columnar storage and SQL queries for large-scale analytics, business intelligence, and data warehousing over large datasets.
Use Firestore when the application needs a flexible document model with automatic scaling and live updates for connected clients. It stores data in documents organized into collections and supports schema flexibility, real-time synchronization, and offline support for use cases such as user profiles, catalogs, chat apps, and collaborative applications.
Regional buckets replicate data synchronously across zones in the region, which supports zero or near-zero recovery time objective and recovery point objective. Dual-region and multi-region buckets store data redundantly in at least two geographic locations with metadata written synchronously across regions and data replicated asynchronously, and turbo replication ensures dual-region objects are replicated across the region pair with a recovery point objective of 15 minutes.
An enterprise is deploying a mission-critical transactional banking application. The application requires a relational database that provides:
Which Google Cloud data management service meets these architectural requirements?
Cloud Storage
Cloud Spanner
Cloud SQL
Cloud Bigtable
Last updated