Associate Data Practitioner

Understand GCP Storage Location Types

Data residency means choosing the specific Google Cloud location where your data is stored at rest. This choice determines how your data is copied for safety (replicated) and how fast it can be accessed. By setting a location, you pick a region for your primary data and keep all related resources inside that geographic area. This decision impacts performance, cost, and meeting legal rules.

In services like Cloud Storage and Cloud SQL, your choice of location type controls data replication. For example, all customer data, including datasets and models, stays within the chosen region and is never moved outside it. This ensures consistent access speed (latency) and helps satisfy regional compliance laws.

You can also use organizational constraints (constraints/gcp.resourceLocations) to enforce company policies. These constraints limit where new resources, like storage buckets, can be created. For instance, you could restrict creation to only us-central1 and europe-west1, preventing accidental use of other regions. This control happens when the resource is created, ensuring all storage complies with your rules.

Align Storage Locations with Redundancy Requirements

Google Cloud's infrastructure is built on regions and zones. A region is a large geographic area, and a zone is a single data center inside that region. A region has at least three separate zones to provide data redundancy. If one zone fails, the others can keep services running.

You need to know the difference between zonal resources and regional resources. Zonal resources, like a single virtual machine, exist in only one zone. If that zone has an outage, those resources become unavailable. Regional resources, like a Cloud Storage bucket set to a regional location, are automatically spread across multiple zones within one region. This provides high availability and protects against a single zone failure.

For protection against larger disasters, you use multi-regions or secondary locations. A multi-region is a broad area like the United States that contains two or more geographic places. Multi-regional locations are often used for secondary data like backups. This ensures your data survives even if an entire primary region is lost.

To design a resilient system, match your storage location to your needs. Your primary data should usually be in the same region as your computing to keep access fast. Your secondary data, like backups, should be in a different region or a multi-region to meet redundancy needs. For example, a Cloud SQL database can run in one region while its backups are stored in a multi-regional location.

Analyze the Advantages and Limitations of Storage Locations

Choosing a storage location involves balancing latency, availability, data durability, and cost. The main types are regions, dual-regions, multi-regions, and zones.

Regions

A region is a specific geographical area. Storing data in one region keeps it close to local users, which reduces access latency. It also allows for quick recovery from a zone failure within that region. However, data stored only in one region is vulnerable to a complete regional outage. Costs can also be higher if you frequently move data between different regions.

Dual-regions

A dual-region combines two distinct geographic regions. This setup greatly improves data durability because your data is copied in two separate places, protecting against an outage in one entire region. It also maintains availability if one region goes down. The trade-offs are potentially slightly higher latency when data needs to sync between the two regions, and generally higher costs than using a single region.

Multi-regions

A multi-region spreads data across multiple regions within a large area like a continent. This maximizes high availability and redundancy, so the failure of one region doesn't stop data access. It's ideal for applications with global users. The main downsides are higher costs due to more infrastructure, and the potential for increased latency from cross-region data synchronization.

Zones

Zones are the subdivisions within a region. Using multiple zones within one region provides localized redundancy, as zones can fail independently. However, zones offer less robust protection than cross-regional setups because all zones are still within the same broader geographic area. Data confined to specific zones might also create latency challenges for applications with users spread out across the world.