All Products
Search
Document Center

Object Storage Service:Bucket overview

Last Updated:Mar 19, 2026

A bucket is the basic container for objects in Object Storage Service (OSS). All objects are stored directly in a bucket — there is no directory hierarchy like a traditional file system.

image

To simulate a folder structure, use a forward slash (/) as a delimiter in object names — for example, folder/file.jpg. The OSS console and OSS Browser display these prefixes as folders.

Key concepts

Region

A region is the physical location of the data center where the bucket resides. The region cannot be changed after the bucket is created.

Storage class

Choose a storage class based on how frequently you access your data. This choice directly affects storage costs.

Storage classKey featuresMinimum storage durationTypical use cases
StandardSupports frequent read and write accessNoneDynamic data for websites and applications
Infrequent AccessLower access frequency30 daysMonitoring data, logs, and enterprise backups
Archive StorageRarely accessed data60 daysLong-term archiving and medical imaging
Cold Archive/Deep Cold ArchiveAlmost never accessed; must be restored before access180 daysCompliance data and historical image archives

If you are unsure which storage class to use, start with Standard. Then use lifecycle rules to automatically transition data to a lower-cost storage class as it ages. For more details, see Storage classes and OSS pricing.

Storage redundancy type

The redundancy type determines how many copies of your data OSS maintains and the resulting reliability level.

Redundancy typeData reliabilityRecommended for
LRS (locally redundant storage)99.999999999% (11 nines)Development and testing environments
ZRS (zone-redundant storage)99.9999999999% (12 nines)Production environments

You can upgrade a bucket from LRS to ZRS, but cannot downgrade from ZRS to LRS. For more details, see Data redundancy types.

Basic operations

Note

An Alibaba Cloud account can create a maximum of 100 buckets per region. A single bucket has no capacity limit.

Access control

OSS enforces bucket access using a layered model. Policies are applied in the following order, from highest to lowest precedence:

  1. Block Public Access — A bucket-level switch that blocks all public access, overriding any ACL or bucket policy that might otherwise grant it. Keep this enabled unless you have a specific requirement for public access. This setting is on by default for all new buckets created in the console.

  2. Bucket policies — Resource-based policies that grant fine-grained access to specific buckets and objects, including access from other Alibaba Cloud accounts or anonymous users.

  3. RAM Policy — User-based policies that control which OSS resources Resource Access Management (RAM) users under your account can access. Use RAM policies to grant least-privilege access to internal users and applications.

  4. Bucket ACL — A simplified permission model with three preset levels: private, public-read, and public-read-write.

Scenario-specific configurations

Ensure data security and compliance

Protect data from unauthorized access and accidental leaks with a layered security approach:

Implement data disaster recovery and backup

Prevent data loss from accidental deletion or regional failures:

  • Prevent accidental deletion: Enable versioning to retain historical versions of objects and restore them at any time.

  • Automate backups: Use the scheduled backup feature to back up your data automatically every day.

Reduce long-term storage costs

For data that is accessed less frequently over time — such as logs and backup files — configure lifecycle rules to automatically transition objects from Standard to Infrequent Access or Archive Storage, and delete them after they expire. This optimizes storage costs without manual intervention.

Accelerate global access

To improve upload and download speeds for a global user base, enable transfer acceleration. This feature routes traffic through a global network of access points to accelerate data transfers across countries and regions. For static websites or frequently accessed small files, combine transfer acceleration with CDN.

Automate data processing and analytics

To trigger downstream workflows automatically when objects are uploaded to OSS, use event notification. When events such as object creation or deletion occur, OSS sends notifications to a destination such as Message Service (MNS) or Function Compute (FC). Use this to build event-driven pipelines for tasks like automated image processing, video and audio processing, or data analytics.