Bucket Locations | Tigris Object Storage Documentation

Overview

A bucket's location defines how and where your object data is stored. You choose a location type when you create a bucket, and it determines the data placement, replication behavior, availability, and consistency model for all objects in that bucket.

Tigris supports four bucket location types:

Location Types

Global

Global is the default location type and is unique to Tigris. Data is distributed globally and automatically stored closest to the request origin. As access patterns change, data migrates to where it's most frequently needed.

How it works:

1. When you upload an object, it is stored in the region nearest to the upload origin.
2. Metadata is replicated to all Tigris regions.
3. When the object is accessed from a different region, a local copy is created in that region automatically.
4. Over time, data migrates to regions where it's accessed most frequently.

Best for: Global applications with distributed users, media delivery, and any use case where you want zero-configuration global performance.

Global is the default location type. If you create a bucket without specifying a location type, it will be Global.

Multi-Region

Multi-region provides the highest availability by maintaining two or more copies of your data across regions within a chosen geography. You select the geography (USA or EUR) when creating the bucket — Tigris automatically selects the regions within that geography where data is stored.

Supported geographies:

Best for: Mission-critical data requiring the highest availability, enterprise workloads where regional failures cannot cause downtime, and applications that need strong consistency globally with geographic redundancy.

Dual-Region

Dual-region provides high availability by maintaining data residency across specific regions of your choice. Unlike multi-region, you have explicit control over exactly which regions store your data.

Region pairing: You can pair any two Tigris regions. Common examples:

Best for: Compliance or policy requirements that mandate data residency in specific regions. If your primary goal is geographic redundancy without a strict two-region constraint, consider multi-region for higher availability, stronger consistency, and better cost-effectiveness.

Single-Region

Single-region stores your data with redundancy across availability zones within a single region. This gives you full control over data residency while maintaining durability within the region.

You can find the list of available regions in the Regions Reference.

Best for: Strict data residency and sovereignty requirements, cost optimization for region-local workloads, and applications where compute and storage are co-located in the same region.

Choosing a Location Type

Decision Guide

For most workloads that need geographic redundancy, we recommend multi-region over dual-region. Multi-region provides higher availability, strong global consistency, and is typically more cost-effective. Choose dual-region only when compliance or policy requires data in specific regions.

Consistency Model Summary

Availability Comparison

Considerations

Performance

• For the lowest latency with zero configuration, use Global. Data automatically moves to where it's accessed most.
• If your workload is concentrated in one geography, Multi-region gives you both performance and the highest availability.
• If compute and storage are co-located, Single-region eliminates cross-region latency entirely.

Compliance and Data Residency

• Single-region gives you the strictest data residency — data never leaves the chosen region.
• Dual-region lets you maintain data residency across more than one region, useful for regulations that allow data within a defined set of jurisdictions (e.g., two EU regions for GDPR compliance).
• Multi-region constrains data to a geography (e.g., Europe), but Tigris selects the specific regions within it.
• Global distributes data based on access patterns — not suitable for strict data residency requirements.

Cost

• Single-region has the lowest storage cost — one region, no cross-region replication.
• Dual-region is billed at the single-region price for each underlying region. For example, a dual-region bucket spanning iad and ord is billed for iad and for ord — effectively doubling your storage cost compared to a single-region bucket.
• Multi-region has a higher per-GB storage price than single-region, but you are billed once — not per underlying region. For most workloads requiring geographic redundancy, multi-region is more cost-effective than dual-region while also providing higher availability and strong global consistency.
• Global is cost-effective for globally accessed data because replication happens on demand rather than eagerly.