Bucket region attributes
Object Storage Service (OSS) buckets are categorized by their region attribute: region-specific or region-agnostic.
Region-specific buckets
Region-specific buckets are suitable for general-purpose OSS use cases. When you create a bucket, select the region-specific attribute and specify the bucket's region, for example, China (Hangzhou). For more information, see Endpoints.
Region-agnostic buckets
Region-agnostic buckets are for scenarios that require only public network access and are not tied to a specific region. When you create a bucket, select the region-agnostic attribute. Data in a region-agnostic bucket is stored in the Chinese mainland, but its exact region is not disclosed. You can access this bucket only through its public endpoint.
Region-agnostic buckets have a fixed region of oss-rg-china-mainland and can be accessed only over the public network. Their public endpoint is fixed to oss-rg-china-mainland.aliyuncs.com. For more information about regions and their corresponding endpoints, see Endpoints.
Feature comparison
"✓" indicates that the feature is supported. "×" indicates that the feature is not supported.
Bucket-level
Operation | Description | Region-specific bucket | Region-agnostic bucket |
To store an object in OSS, you must first create a bucket. A bucket has various configuration properties, including its region, access control list (ACL), and other metadata. When you create a bucket, select a suitable region based on your requirements for latency, cost, and compliance. For more information, see How to select an OSS region. For a list of supported regions, see Regions and endpoints. | ✓ | ✓ | |
You can list all or some of the buckets in a region by setting different listing conditions. | ✓ | ✓ | |
You can delete a bucket that you no longer need. | ✓ | ✓ | |
Use the GetBucketLocation API operation to get a bucket's region, which corresponds to the physical location of its data center. | ✓ | ✓ | |
A resource group is a resource-based method for permission management. You can group buckets based on business requirements and configure different access permissions for each resource group. | ✓ | × | |
In Requester Pays mode, the requester pays for the fees for accessing data in a bucket, while the bucket owner pays only for storage. You can enable this feature if you want to share data without incurring data access fees. | ✓ | × | |
You can use tags to categorize buckets for easier management, such as listing buckets that have specific tags or configuring access permissions for buckets with specific tags. | ✓ | × | |
As internet services evolve, the demand for higher data throughput increases. To meet this requirement, OSS provides an accelerator feature to cache hot objects and deliver high-performance, high-throughput data access services. | ✓ | × | |
You can set an access control list (ACL) when you create a bucket or modify the ACL of an existing bucket based on your business requirements. Only the bucket owner can perform this operation. | ✓ | ✓ | |
A Bucket Policy is an authorization policy for buckets in OSS. You can use a Bucket Policy to authorize other users to access your specified OSS resources. | ✓ | ✓ | |
Versioning is a bucket-level data protection feature. After you enable versioning, overwrite and delete operations on your data are saved as historical versions. If you accidentally overwrite or delete an object, you can restore the object to any of its previous versions. | ✓ | ✓ | |
OSS allows you to configure hotlinking protection for a bucket. This feature prevents your OSS resources from being used by others by adding referrers to an allowlist. | ✓ | ✓ | |
Cross-Origin Resource Sharing (CORS) is a standard cross-origin solution provided by HTML5. CORS allows web application servers to control cross-origin access and ensures the security of cross-origin data transfers. | ✓ | ✓ | |
Sensitive data protection is a native OSS service that identifies, classifies, grades, and protects sensitive data in buckets. This helps you meet the compliance requirements of data security and personal information protection regulations. | ✓ | × | |
OSS supports the WORM feature, which allows you to store and use data in a way that prevents it from being deleted or modified. This feature meets the compliance requirements of the U.S. Securities and Exchange Commission (SEC) and the Financial Industry Regulatory Authority (FINRA). | ✓ | ✓ | |
When you upload an object to a bucket configured with server-side encryption, OSS encrypts the object before storing it. When you download the object by using a GetObject request, OSS automatically decrypts it and returns it to you. The x-oss-server-side-encryption header is included in the response to indicate that the object was encrypted on the server side. | ✓ | Only SSE-OSS is supported. SSE-KMS is not supported. | |
Client-side encryption refers to encrypting objects locally before sending them to OSS. This topic describes the methods and processes for client-side encryption. | ✓ | ✓ | |
OSS allows you to enable direct read for Archive objects in a bucket. After this feature is enabled, you can directly access objects of the Archive storage class without restoring them first. Compared with restoring Archive objects before reading them, direct read offers shorter data retrieval times but incurs higher costs. | ✓ | × | |
OSS uses a multi-availability zone (AZ) data redundancy mechanism to store copies of your data across multiple AZs in the same region. If an AZ becomes unavailable, your data remains accessible. ZRS is designed to provide data durability of 99.9999999999% (12 nines) and service availability of 99.995%. | ✓ | × | |
OSS lets you convert the storage redundancy type of a bucket from locally redundant storage (LRS) to zone-redundant storage (ZRS) to provide resilience against AZ-level failures. | ✓ | × | |
OSS Anti-DDoS is a DDoS attack proxy protection service that OSS provides in conjunction with Anti-DDoS. When a protected bucket is subject to a high-volume attack, OSS Anti-DDoS redirects the attack traffic to an Anti-DDoS cluster for scrubbing and forwards normal traffic back to the target bucket to ensure business continuity. | ✓ | × | |
Cross-region replication (CRR) automatically and asynchronously (in near real time) replicates objects between buckets in different OSS data centers. Operations such as creating, updating, and deleting objects are replicated from a source bucket to a destination bucket in another region. | ✓ | CRR can be configured only between a region-agnostic bucket and a bucket in a Chinese mainland region. | |
Same-region replication (SRR) automatically and asynchronously (in near real time) replicates object operations such as creation, updates, and deletions from a source bucket to a destination bucket in the same region. | ✓ | To replicate data between two region-agnostic buckets, you must use cross-region replication (CRR). | |
Replication Time Control (RTC) for data replication helps you meet compliance or business requirements for cross-region data replication. When RTC is enabled, OSS replicates most of the objects that you upload to OSS within seconds and 99.99% of the objects within 10 minutes. RTC also provides near real-time monitoring of data replication, allowing you to view various metrics of replication tasks. | ✓ | × | |
You can use the scheduled backup feature of OSS to periodically back up objects in a bucket to Cloud Backup. If your objects are accidentally lost, you can restore them from Cloud Backup. | ✓ | × | |
Lifecycle rules allow you to periodically transition objects in a bucket to the Infrequent Access (IA), Archive, Cold Archive, or Deep Cold Archive storage class, or delete expired objects and parts to save on storage costs. | ✓ | Only deletion is supported. Storage class transitions are not supported. | |
You can use a lifecycle rule based on the last access time to automatically monitor data access patterns, identify cold data, and then transition the identified cold data to a different storage class. This enables tiered storage for hot and cold data and reduces storage costs. | ✓ | × | |
You can use the inventory feature to obtain information about specified objects in a bucket, such as the object count, size, storage class, and encryption status. We recommend that you use the inventory feature instead of the GetBucket (ListObjects) API operation to list a large number of objects. | ✓ | × | |
A static website consists of static content, such as client-side scripts like JavaScript. You can use the static website hosting feature to host your static website in an OSS bucket and access it through the bucket's endpoint. | ✓ | ✓ | |
After you configure a back-to-origin rule, if a requester tries to access an object that does not exist in the bucket, OSS fetches the object from the origin site specified in the rule. After fetching the object, OSS returns it to the requester and stores it in the bucket. | ✓ | ✓ | |
OSS Transfer Acceleration uses globally distributed cloud data centers to provide an end-to-end acceleration solution for uploads and downloads. When users from anywhere in the world access your bucket, their requests are directed to the nearest access point through intelligent routing and transmitted over optimized networks and protocols. | ✓ | × | |
After you upload an object to a bucket, OSS automatically generates an object URL that you can use to access the object. If you want to access these objects through a custom domain name, you must map the custom domain name to the bucket and add a CNAME record. | ✓ | ✓ | |
You can process image objects stored in OSS by including image processing parameters in a GetObject request. For example, you can add image watermarks or convert image formats. | ✓ | ✓ | |
OSS supports capturing a snapshot from a video at a specified time. | ✓ | ✓ | |
OSS is deeply integrated with Intelligent Media Management (IMM) and supports a wide range of data analysis and processing operations, such as document preview, document format conversion, face recognition, image analysis, and QR code recognition. | ✓ | × | |
OSS provides an automatic decompression feature for ZIP files. After you configure a decompression rule, ZIP files that you upload to the specified path in OSS are automatically decompressed. | ✓ | × | |
When you configure event notifications for a bucket, you can specify the objects that you want to monitor. When a specified event occurs on these objects, you receive an immediate notification. | ✓ | × | |
The OSS monitoring service provides monitoring metrics for basic system operations, performance, and metering. It also offers a custom alert service to help you track requests, analyze usage, and identify business trends, allowing you to promptly diagnose and resolve system issues. | ✓ | ✓ | |
Accessing OSS generates a large number of access logs. You can use the server access logging feature to generate log files hourly based on fixed naming rules and write them to a bucket that you specify. You can analyze the stored logs by using services such as Alibaba Cloud Log Service or by building a Spark cluster. | ✓ | Logs can be saved only to the current bucket. | |
After you enable real-time log query, you can track access requests to your bucket. This helps you perform operation audits, access statistics, and incident analysis, which improves your work efficiency and helps you make better data-driven decisions. | ✓ | × | |
The OSS-HDFS service (JindoFS service) is a cloud-native data lake storage product. It uses unified metadata management capabilities to provide full POSIX support and is fully compatible with HDFS file system interfaces. This makes it ideal for data lake computing scenarios in big data and AI. | ✓ | × | |
OSS ON CloudBox provides local storage, local access, and local processing capabilities for unstructured data in a CloudBox. You can create buckets in OSS ON CloudBox and use the same APIs and SDKs as in the public cloud to access OSS within the CloudBox. | ✓ | × |
Object-level
Operation | Description | Region-specific bucket | Region-agnostic bucket |
A simple upload uses the PutObject API operation to upload a single object that is smaller than 5 GB. It is suitable for scenarios where the upload can be completed in a single HTTP request. | ✓ | ✓ | |
Multipart upload splits an object into multiple parts, uploads them separately, and then calls the CompleteMultipartUpload operation to combine these parts into a single object. | ✓ | ✓ | |
Resumable upload uses a checkpoint file to record breakpoints. If an upload is interrupted by a network issue or program crash, it can resume from the last recorded breakpoint. | ✓ | ✓ | |
OSS can send a callback to an application server after an object is uploaded. To implement this, include the appropriate callback parameters in the request sent to OSS. | ✓ | ✓ | |
Client-side direct upload lets clients upload objects directly to OSS. Compared to server-side proxy uploads, this method avoids file transfers through your business server, which improves upload speed and saves server resources. | ✓ | ✓ | |
A form upload uses the PostObject API operation to upload an object. The uploaded object cannot exceed 5 GB in size. | ✓ | ✓ | |
An append upload adds content directly to the end of an existing appendable object. | ✓ | ✓ | |
OSS supports ingesting H.264-encoded video streams and AAC-encoded audio streams by using RTMP. The ingested audio and video data can be played on demand. It can also be used for live streaming in applications where latency is not a critical concern. | ✓ | × | |
A simple download uses the GetObject API operation to download an object. It is suitable for scenarios where the download can be completed in a single HTTP request. | ✓ | ✓ | |
OSS provides the ability to download an object starting from a specified position. This allows you to download large objects in multiple parts. If a download is interrupted, you can resume it from the last completed position. | ✓ | ✓ | |
You can authorize third parties to download an object by using temporary access credentials and a signed URL, which avoids exposing your account's AccessKey pair. | ✓ | ✓ | |
Objects in a bucket are listed alphabetically by default. You can list all objects, objects with a specific prefix, or a specific number of objects based on your needs. | ✓ | ✓ | |
This operation replicates an object from a source bucket to a destination bucket in the same region. | ✓ | ✓ | |
You can rename an object within the same bucket. | ✓ | ✓ | |
After you upload an object to a bucket, you can share it with a third party for preview or download by using its URL. | ✓ | ✓ | |
When you have a large number of objects in a bucket, you can quickly search for and locate a target object by specifying an object name prefix. | ✓ | ✓ | |
Objects in the Cold Archive or Deep Cold Archive storage class must be restored before they can be read. | ✓ | ✓ | |
You can delete objects that you no longer need from a bucket in several ways. | ✓ | ✓ | |
OSS supports object tagging to categorize objects in a bucket. You can then configure lifecycle rules and access permissions for objects with the same tags. | ✓ | ✓ | |
Symbolic links provide quick access to frequently used objects in a bucket. After a symbolic link is configured, you can use it like a shortcut in Windows to quickly open an object. | ✓ | ✓ | |
Object metadata describes the properties of an object. It includes standard HTTP headers and user-defined metadata. You can set HTTP headers to define custom HTTP request policies, such as caching or forced download policies. You can also set user-defined metadata to identify the purpose or properties of an object. | ✓ | ✓ | |
Accessing objects in OSS can consume significant bandwidth, which may affect other applications on clients where traffic is difficult to control. To avoid this issue, you can use the single-link speed limit feature in OSS to control traffic during operations like uploads and downloads, ensuring network bandwidth for other applications. | ✓ | ✓ | |
You can execute SQL statements on a target object by using SelectObject and receive the execution results. | ✓ | × | |
If a multipart upload is not completed, the uploaded parts remain in the bucket and incur storage costs. You can delete these incomplete parts to save space and reduce costs. | ✓ | ✓ | |
Unlike the hierarchical structure of traditional file systems, OSS uses a flat structure to store data. All data is stored in buckets as objects. To simplify management, you can organize these objects into a directory-like hierarchy. | ✓ | ✓ | |
Data indexing is an object metadata indexing service provided by OSS. You can customize indexing conditions based on object metadata to quickly retrieve a list of objects. Data indexing helps you better understand and manage your data structure, making it easier to query, analyze, and manage objects. | ✓ | × |
FAQ
Data migration to region-agnostic buckets
Yes. You can use cross-region replication (CRR) to migrate data from a region-specific bucket to a region-agnostic bucket. For more information, see Cross-region replication.
You cannot migrate data from a region-specific bucket to a region-agnostic bucket if the source bucket is encrypted with SSE-KMS.